You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@iceberg.apache.org by gi...@apache.org on 2026/09/14 18:21:46 UTC

(iceberg-rust) branch gh-pages updated: deploy: f11a20d0ba9b6c168839e234f3817b0e0d75ce4b

This is an automated email from the ASF dual-hosted git repository.

github-actions[bot] pushed a commit to branch gh-pages
in repository https://gitbox.apache.org/repos/asf/iceberg-rust.git


The following commit(s) were added to refs/heads/gh-pages by this push:
     new 529e10a57 deploy: f11a20d0ba9b6c168839e234f3817b0e0d75ce4b
529e10a57 is described below

commit 529e10a57cf77b6cbb54a62384d3a5b077cdba9e
Author: github-merge-queue[bot] <gi...@users.noreply.github.com>
AuthorDate: Mon Sep 14 18:21:40 2026 +0000

    deploy: f11a20d0ba9b6c168839e234f3817b0e0d75ce4b
---
 .../constant.PROJECTED_PARTITION_VALUE_COLUMN.html |   2 +-
 .../record_batch_partition_splitter/index.html     |   2 +-
 .../struct.RecordBatchPartitionSplitter.html       |  10 +-
 .../path/{c9deffff14.js => 33779bc049.js}          |   2 +-
 api/search.index/root.js                           |   2 +-
 .../arrow/record_batch_partition_splitter.rs.html  | 914 +++++++++++----------
 6 files changed, 467 insertions(+), 465 deletions(-)

diff --git a/api/iceberg/arrow/record_batch_partition_splitter/constant.PROJECTED_PARTITION_VALUE_COLUMN.html b/api/iceberg/arrow/record_batch_partition_splitter/constant.PROJECTED_PARTITION_VALUE_COLUMN.html
index e3761b1ab..1d9e4a433 100644
--- a/api/iceberg/arrow/record_batch_partition_splitter/constant.PROJECTED_PARTITION_VALUE_COLUMN.html
+++ b/api/iceberg/arrow/record_batch_partition_splitter/constant.PROJECTED_PARTITION_VALUE_COLUMN.html
@@ -1,2 +1,2 @@
-<!DOCTYPE html><html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><meta name="generator" content="rustdoc"><meta name="description" content="Column name for the projected partition values struct"><title>PROJECTED_PARTITION_VALUE_COLUMN in iceberg::arrow::record_batch_partition_splitter - Rust</title><script>if(window.location.protocol!=="file:")document.head.insertAdjacentHTML("beforeend","SourceSerif4-Regular-6b053e98.ttf.wo [...]
+<!DOCTYPE html><html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><meta name="generator" content="rustdoc"><meta name="description" content="Column name for the projected partition values struct"><title>PROJECTED_PARTITION_VALUE_COLUMN in iceberg::arrow::record_batch_partition_splitter - Rust</title><script>if(window.location.protocol!=="file:")document.head.insertAdjacentHTML("beforeend","SourceSerif4-Regular-6b053e98.ttf.wo [...]
 </div></details></section></div></main></body></html>
\ No newline at end of file
diff --git a/api/iceberg/arrow/record_batch_partition_splitter/index.html b/api/iceberg/arrow/record_batch_partition_splitter/index.html
index f29bed44d..bca145da8 100644
--- a/api/iceberg/arrow/record_batch_partition_splitter/index.html
+++ b/api/iceberg/arrow/record_batch_partition_splitter/index.html
@@ -1,2 +1,2 @@
-<!DOCTYPE html><html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><meta name="generator" content="rustdoc"><meta name="description" content="Record batch partition splitter for partitioned tables"><title>iceberg::arrow::record_batch_partition_splitter - Rust</title><script>if(window.location.protocol!=="file:")document.head.insertAdjacentHTML("beforeend","SourceSerif4-Regular-6b053e98.ttf.woff2,FiraSans-Italic-81dc35de.woff2, [...]
+<!DOCTYPE html><html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><meta name="generator" content="rustdoc"><meta name="description" content="Record batch partition splitter for partitioned tables"><title>iceberg::arrow::record_batch_partition_splitter - Rust</title><script>if(window.location.protocol!=="file:")document.head.insertAdjacentHTML("beforeend","SourceSerif4-Regular-6b053e98.ttf.woff2,FiraSans-Italic-81dc35de.woff2, [...]
 </div></details><h2 id="structs" class="section-header">Structs<a href="#structs" class="anchor">§</a></h2><dl class="item-table"><dt><a class="struct" href="struct.RecordBatchPartitionSplitter.html" title="struct iceberg::arrow::record_batch_partition_splitter::RecordBatchPartitionSplitter">Record<wbr>Batch<wbr>Partition<wbr>Splitter</a></dt><dd>The splitter used to split the record batch into multiple record batches by the partition spec.</dd></dl><h2 id="constants" class="section-head [...]
\ No newline at end of file
diff --git a/api/iceberg/arrow/record_batch_partition_splitter/struct.RecordBatchPartitionSplitter.html b/api/iceberg/arrow/record_batch_partition_splitter/struct.RecordBatchPartitionSplitter.html
index 41b6c40cc..54080eef1 100644
--- a/api/iceberg/arrow/record_batch_partition_splitter/struct.RecordBatchPartitionSplitter.html
+++ b/api/iceberg/arrow/record_batch_partition_splitter/struct.RecordBatchPartitionSplitter.html
@@ -1,4 +1,4 @@
-<!DOCTYPE html><html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><meta name="generator" content="rustdoc"><meta name="description" content="The splitter used to split the record batch into multiple record batches by the partition spec."><title>RecordBatchPartitionSplitter in iceberg::arrow::record_batch_partition_splitter - Rust</title><script>if(window.location.protocol!=="file:")document.head.insertAdjacentHTML("beforeend" [...]
+<!DOCTYPE html><html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><meta name="generator" content="rustdoc"><meta name="description" content="The splitter used to split the record batch into multiple record batches by the partition spec."><title>RecordBatchPartitionSplitter in iceberg::arrow::record_batch_partition_splitter - Rust</title><script>if(window.location.protocol!=="file:")document.head.insertAdjacentHTML("beforeend" [...]
 <ol>
 <li>It will project and transform the input record batch based on the partition spec, get the partitioned record batch.</li>
 <li>Split the input record batch into multiple record batches based on the partitioned record batch.</li>
@@ -9,7 +9,7 @@
 <li><strong>Computed mode</strong> (<code>calculator</code> is <code>Some</code>): Computes partition values from source columns using transforms</li>
 <li><strong>Pre-computed mode</strong> (<code>calculator</code> is <code>None</code>): Expects a <code>_partition</code> column in the input batch</li>
 </ul>
-</div></details><h2 id="implementations" class="section-header">Implementations<a href="#implementations" class="anchor">§</a></h2><div id="implementations-list"><details class="toggle implementors-toggle" open><summary><section id="impl-RecordBatchPartitionSplitter" class="impl"><a class="src rightside" href="../../../src/iceberg/arrow/record_batch_partition_splitter.rs.html#48-213">Source</a><a href="#impl-RecordBatchPartitionSplitter" class="anchor">§</a><h3 class="code-header">impl < [...]
+</div></details><h2 id="implementations" class="section-header">Implementations<a href="#implementations" class="anchor">§</a></h2><div id="implementations-list"><details class="toggle implementors-toggle" open><summary><section id="impl-RecordBatchPartitionSplitter" class="impl"><a class="src rightside" href="../../../src/iceberg/arrow/record_batch_partition_splitter.rs.html#49-215">Source</a><a href="#impl-RecordBatchPartitionSplitter" class="anchor">§</a><h3 class="code-header">impl < [...]
     iceberg_schema: <a class="type" href="../../spec/type.SchemaRef.html" title="type iceberg::spec::SchemaRef">SchemaRef</a>,
     partition_spec: <a class="type" href="../../spec/type.PartitionSpecRef.html" title="type iceberg::spec::PartitionSpecRef">PartitionSpecRef</a>,
     calculator: <a class="enum" href="https://doc.rust-lang.org/nightly/core/option/enum.Option.html" title="enum core::option::Option">Option</a>&lt;<a class="struct" href="../partition_value_calculator/struct.PartitionValueCalculator.html" title="struct iceberg::arrow::partition_value_calculator::PartitionValueCalculator">PartitionValueCalculator</a>&gt;,
@@ -27,7 +27,7 @@
 </ul>
 <h5 id="returns"><a class="doc-anchor" href="#returns">§</a>Returns</h5>
 <p>Returns a new <code>RecordBatchPartitionSplitter</code> instance or an error if initialization fails.</p>
-</div></details><details class="toggle method-toggle" open><summary><section id="method.try_new_with_computed_values" class="method"><a class="src rightside" href="../../../src/iceberg/arrow/record_batch_partition_splitter.rs.html#90-96">Source</a><h4 class="code-header">pub fn <a href="#method.try_new_with_computed_values" class="fn">try_new_with_computed_values</a>(
+</div></details><details class="toggle method-toggle" open><summary><section id="method.try_new_with_computed_values" class="method"><a class="src rightside" href="../../../src/iceberg/arrow/record_batch_partition_splitter.rs.html#91-97">Source</a><h4 class="code-header">pub fn <a href="#method.try_new_with_computed_values" class="fn">try_new_with_computed_values</a>(
     iceberg_schema: <a class="type" href="../../spec/type.SchemaRef.html" title="type iceberg::spec::SchemaRef">SchemaRef</a>,
     partition_spec: <a class="type" href="../../spec/type.PartitionSpecRef.html" title="type iceberg::spec::PartitionSpecRef">PartitionSpecRef</a>,
 ) -&gt; <a class="type" href="../../type.Result.html" title="type iceberg::Result">Result</a>&lt;Self&gt;</h4></section></summary><div class="docblock"><p>Create a new RecordBatchPartitionSplitter with computed partition values.</p>
@@ -40,7 +40,7 @@ to compute partition values from source columns.</p>
 </ul>
 <h5 id="returns-1"><a class="doc-anchor" href="#returns-1">§</a>Returns</h5>
 <p>Returns a new <code>RecordBatchPartitionSplitter</code> instance or an error if initialization fails.</p>
-</div></details><details class="toggle method-toggle" open><summary><section id="method.try_new_with_precomputed_values" class="method"><a class="src rightside" href="../../../src/iceberg/arrow/record_batch_partition_splitter.rs.html#111-116">Source</a><h4 class="code-header">pub fn <a href="#method.try_new_with_precomputed_values" class="fn">try_new_with_precomputed_values</a>(
+</div></details><details class="toggle method-toggle" open><summary><section id="method.try_new_with_precomputed_values" class="method"><a class="src rightside" href="../../../src/iceberg/arrow/record_batch_partition_splitter.rs.html#112-117">Source</a><h4 class="code-header">pub fn <a href="#method.try_new_with_precomputed_values" class="fn">try_new_with_precomputed_values</a>(
     iceberg_schema: <a class="type" href="../../spec/type.SchemaRef.html" title="type iceberg::spec::SchemaRef">SchemaRef</a>,
     partition_spec: <a class="type" href="../../spec/type.PartitionSpecRef.html" title="type iceberg::spec::PartitionSpecRef">PartitionSpecRef</a>,
 ) -&gt; <a class="type" href="../../type.Result.html" title="type iceberg::Result">Result</a>&lt;Self&gt;</h4></section></summary><div class="docblock"><p>Create a new RecordBatchPartitionSplitter expecting pre-computed partition values.</p>
@@ -53,7 +53,7 @@ column in the input batches.</p>
 </ul>
 <h5 id="returns-2"><a class="doc-anchor" href="#returns-2">§</a>Returns</h5>
 <p>Returns a new <code>RecordBatchPartitionSplitter</code> instance or an error if initialization fails.</p>
-</div></details><details class="toggle method-toggle" open><summary><section id="method.split" class="method"><a class="src rightside" href="../../../src/iceberg/arrow/record_batch_partition_splitter.rs.html#119-212">Source</a><h4 class="code-header">pub fn <a href="#method.split" class="fn">split</a>(
+</div></details><details class="toggle method-toggle" open><summary><section id="method.split" class="method"><a class="src rightside" href="../../../src/iceberg/arrow/record_batch_partition_splitter.rs.html#120-214">Source</a><h4 class="code-header">pub fn <a href="#method.split" class="fn">split</a>(
     &amp;self,
     batch: &amp;RecordBatch,
 ) -&gt; <a class="type" href="../../type.Result.html" title="type iceberg::Result">Result</a>&lt;<a class="struct" href="https://doc.rust-lang.org/nightly/alloc/vec/struct.Vec.html" title="struct alloc::vec::Vec">Vec</a>&lt;(<a class="struct" href="../../spec/struct.PartitionKey.html" title="struct iceberg::spec::PartitionKey">PartitionKey</a>, RecordBatch)&gt;&gt;</h4></section></summary><div class="docblock"><p>Split the record batch into multiple record batches based on the partition  [...]
diff --git a/api/search.index/path/c9deffff14.js b/api/search.index/path/33779bc049.js
similarity index 62%
rename from api/search.index/path/c9deffff14.js
rename to api/search.index/path/33779bc049.js
index a8666d884..c234e4a4a 100644
--- a/api/search.index/path/c9deffff14.js
+++ b/api/search.index/path/33779bc049.js
@@ -1 +1 @@
-rd_("g[17,\"\"]000f[1,\"\"]0Aa[10,\"bnum::cast\"]Bn[15,\"iceberg_storage_opendal::OpenDalStorage\"]Eg[5,\"tower_http::follow_redirect::policy::or\",\"tower_http::follow_redirect::policy::or\"]1Ce[15,\"iceberg_storage_opendal::OpenDalStorageFactory\"]4Bb[2,\"iceberg::io::storage::config\"]m[2,\"iceberg\"]Ai[2,\"iceberg::encryption\"]2Am[2,\"iceberg_storage_opendal\"]8Ei[5,\"tower_http::follow_redirect::policy::and\",\"tower_http::follow_redirect::policy::and\"]A`[10,\"core::any\"]Ao[5,\"a [...]
\ No newline at end of file
+rd_("g[17,\"\"]000f[1,\"\"]0Aa[10,\"bnum::cast\"]Bn[15,\"iceberg_storage_opendal::OpenDalStorage\"]Eg[5,\"tower_http::follow_redirect::policy::or\",\"tower_http::follow_redirect::policy::or\"]1Ce[15,\"iceberg_storage_opendal::OpenDalStorageFactory\"]4Bb[2,\"iceberg::io::storage::config\"]m[2,\"iceberg\"]Ai[2,\"iceberg::encryption\"]2Am[2,\"iceberg_storage_opendal\"]8Ei[5,\"tower_http::follow_redirect::policy::and\",\"tower_http::follow_redirect::policy::and\"]A`[10,\"core::any\"]Ao[5,\"a [...]
\ No newline at end of file
diff --git a/api/search.index/root.js b/api/search.index/root.js
index 6eb732e9e..f4e4168b5 100644
--- a/api/search.index/root.js
+++ b/api/search.index/root.js
@@ -1 +1 @@
-rr_('{"normalizedName":{"I":"VYLAAAAUexSTAVsH1wxzD5oPJQHAAAAwrAObBAEFNP8EdQ8lAcAAABB1DzZ1F8giRQHAAAAS2QxypQGsA5sEAQV1/MBEAgAAQVsH1wxzD5oPEP8EEXQPHwAgACEAlAFVEScgZiIVAcAAACFMEfAcbqYBBQLAAAATFCAQ7xxhcwEBAcUAABxTZDswAAABAAD0AAMACwUAACIFAACXCfIABQLDAACnMCKiZSFkc0UBwgAAFR4bY44KjwqQCsoKFQHAAAAS2QxypQEUAcJ0ZAAAFmVmYXVsdMAcZCUFdhsAAf4IThwAaKgkpTrEAAAggXMA0kEbAAC+FQHRAAAW0iJkqCIVAdYAABbgImTRIjHAAAAwEfIe8x7zgQJmcGUBwwAAFtoiZNYhYyKXIpgivSLBIkUBwwAAFtQiZNgapx5HIrAiIwACxAAAITZhbvQsHAAApQFWBDcA+4YJAtA [...]
\ No newline at end of file
+rr_('{"normalizedName":{"I":"VYLAAAAUexSTAVsH1wxzD5oPJQHAAAAwrAObBAEFNP8EdQ8lAcAAABB1DzZ1F8giRQHAAAAS2QxypQGsA5sEAQV1/MBEAgAAQVsH1wxzD5oPEP8EEXQPHwAgACEAlAFVEScgZiIVAcAAACFMEfAcbqYBBQLAAAATFCAQ7xxhcwEBAcUAABxTZDswAAABAAD0AAMACwUAACIFAACXCfIABQLDAACnMCKiZSFkc0UBwgAAFR4bY44KjwqQCsoKFQHAAAAS2QxypQEUAcJ0ZAAAFmVmYXVsdMAcZCUFdhsAAf4IThwAaKgkpTrEAAAggXMA0kEbAAC+FQHRAAAW0iJkqCIVAdYAABbgImTRIjHAAAAwEfIe8x7zgQJmcGUBwwAAFtoiZNYhYyKXIpgivSLBIkUBwwAAFtQiZNgapx5HIrAiIwACxAAAITZhbvQsHAAApQFWBDcA+4YJAtA [...]
\ No newline at end of file
diff --git a/api/src/iceberg/arrow/record_batch_partition_splitter.rs.html b/api/src/iceberg/arrow/record_batch_partition_splitter.rs.html
index 8c475b5e6..d5c18052e 100644
--- a/api/src/iceberg/arrow/record_batch_partition_splitter.rs.html
+++ b/api/src/iceberg/arrow/record_batch_partition_splitter.rs.html
@@ -19,463 +19,465 @@
 <a href=#19 id=19 data-nosnippet>19</a><span class="kw">use </span>std::sync::Arc;
 <a href=#20 id=20 data-nosnippet>20</a>
 <a href=#21 id=21 data-nosnippet>21</a><span class="kw">use </span>arrow_array::{ArrayRef, BooleanArray, RecordBatch, StructArray};
-<a href=#22 id=22 data-nosnippet>22</a><span class="kw">use </span>arrow_select::filter::filter_record_batch;
-<a href=#23 id=23 data-nosnippet>23</a>
-<a href=#24 id=24 data-nosnippet>24</a><span class="kw">use </span><span class="kw">super</span>::arrow_struct_to_literal;
-<a href=#25 id=25 data-nosnippet>25</a><span class="kw">use </span><span class="kw">super</span>::partition_value_calculator::PartitionValueCalculator;
-<a href=#26 id=26 data-nosnippet>26</a><span class="kw">use </span><span class="kw">crate</span>::spec::{Literal, PartitionKey, PartitionSpecRef, SchemaRef, StructType};
-<a href=#27 id=27 data-nosnippet>27</a><span class="kw">use </span>crate::{Error, ErrorKind, <span class="prelude-ty">Result</span>};
-<a href=#28 id=28 data-nosnippet>28</a>
-<a href=#29 id=29 data-nosnippet>29</a><span class="doccomment">/// Column name for the projected partition values struct
-<a href=#30 id=30 data-nosnippet>30</a></span><span class="kw">pub const </span>PROJECTED_PARTITION_VALUE_COLUMN: <span class="kw-2">&amp;</span>str = <span class="string">"_partition"</span>;
-<a href=#31 id=31 data-nosnippet>31</a>
-<a href=#32 id=32 data-nosnippet>32</a><span class="doccomment">/// The splitter used to split the record batch into multiple record batches by the partition spec.
-<a href=#33 id=33 data-nosnippet>33</a>/// 1. It will project and transform the input record batch based on the partition spec, get the partitioned record batch.
-<a href=#34 id=34 data-nosnippet>34</a>/// 2. Split the input record batch into multiple record batches based on the partitioned record batch.
-<a href=#35 id=35 data-nosnippet>35</a>///
-<a href=#36 id=36 data-nosnippet>36</a>/// # Partition Value Modes
-<a href=#37 id=37 data-nosnippet>37</a>///
-<a href=#38 id=38 data-nosnippet>38</a>/// The splitter supports two modes for obtaining partition values:
-<a href=#39 id=39 data-nosnippet>39</a>/// - **Computed mode** (`calculator` is `Some`): Computes partition values from source columns using transforms
-<a href=#40 id=40 data-nosnippet>40</a>/// - **Pre-computed mode** (`calculator` is `None`): Expects a `_partition` column in the input batch
-<a href=#41 id=41 data-nosnippet>41</a></span><span class="kw">pub struct </span>RecordBatchPartitionSplitter {
-<a href=#42 id=42 data-nosnippet>42</a>    schema: SchemaRef,
-<a href=#43 id=43 data-nosnippet>43</a>    partition_spec: PartitionSpecRef,
-<a href=#44 id=44 data-nosnippet>44</a>    calculator: <span class="prelude-ty">Option</span>&lt;PartitionValueCalculator&gt;,
-<a href=#45 id=45 data-nosnippet>45</a>    partition_type: StructType,
-<a href=#46 id=46 data-nosnippet>46</a>}
-<a href=#47 id=47 data-nosnippet>47</a>
-<a href=#48 id=48 data-nosnippet>48</a><span class="kw">impl </span>RecordBatchPartitionSplitter {
-<a href=#49 id=49 data-nosnippet>49</a>    <span class="doccomment">/// Create a new RecordBatchPartitionSplitter.
-<a href=#50 id=50 data-nosnippet>50</a>    ///
-<a href=#51 id=51 data-nosnippet>51</a>    /// # Arguments
-<a href=#52 id=52 data-nosnippet>52</a>    ///
-<a href=#53 id=53 data-nosnippet>53</a>    /// * `iceberg_schema` - The Iceberg schema reference
-<a href=#54 id=54 data-nosnippet>54</a>    /// * `partition_spec` - The partition specification reference
-<a href=#55 id=55 data-nosnippet>55</a>    /// * `calculator` - Optional calculator for computing partition values from source columns.
-<a href=#56 id=56 data-nosnippet>56</a>    ///   - `Some(calculator)`: Compute partition values from source columns using transforms
-<a href=#57 id=57 data-nosnippet>57</a>    ///   - `None`: Expect a pre-computed `_partition` column in the input batch
-<a href=#58 id=58 data-nosnippet>58</a>    ///
-<a href=#59 id=59 data-nosnippet>59</a>    /// # Returns
-<a href=#60 id=60 data-nosnippet>60</a>    ///
-<a href=#61 id=61 data-nosnippet>61</a>    /// Returns a new `RecordBatchPartitionSplitter` instance or an error if initialization fails.
-<a href=#62 id=62 data-nosnippet>62</a>    </span><span class="kw">pub fn </span>try_new(
-<a href=#63 id=63 data-nosnippet>63</a>        iceberg_schema: SchemaRef,
-<a href=#64 id=64 data-nosnippet>64</a>        partition_spec: PartitionSpecRef,
-<a href=#65 id=65 data-nosnippet>65</a>        calculator: <span class="prelude-ty">Option</span>&lt;PartitionValueCalculator&gt;,
-<a href=#66 id=66 data-nosnippet>66</a>    ) -&gt; <span class="prelude-ty">Result</span>&lt;<span class="self">Self</span>&gt; {
-<a href=#67 id=67 data-nosnippet>67</a>        <span class="kw">let </span>partition_type = partition_spec.partition_type(<span class="kw-2">&amp;</span>iceberg_schema)<span class="question-mark">?</span>;
-<a href=#68 id=68 data-nosnippet>68</a>
-<a href=#69 id=69 data-nosnippet>69</a>        <span class="prelude-val">Ok</span>(<span class="self">Self </span>{
-<a href=#70 id=70 data-nosnippet>70</a>            schema: iceberg_schema,
-<a href=#71 id=71 data-nosnippet>71</a>            partition_spec,
-<a href=#72 id=72 data-nosnippet>72</a>            calculator,
-<a href=#73 id=73 data-nosnippet>73</a>            partition_type,
-<a href=#74 id=74 data-nosnippet>74</a>        })
-<a href=#75 id=75 data-nosnippet>75</a>    }
-<a href=#76 id=76 data-nosnippet>76</a>
-<a href=#77 id=77 data-nosnippet>77</a>    <span class="doccomment">/// Create a new RecordBatchPartitionSplitter with computed partition values.
-<a href=#78 id=78 data-nosnippet>78</a>    ///
-<a href=#79 id=79 data-nosnippet>79</a>    /// This is a convenience method that creates a calculator and initializes the splitter
-<a href=#80 id=80 data-nosnippet>80</a>    /// to compute partition values from source columns.
-<a href=#81 id=81 data-nosnippet>81</a>    ///
-<a href=#82 id=82 data-nosnippet>82</a>    /// # Arguments
-<a href=#83 id=83 data-nosnippet>83</a>    ///
-<a href=#84 id=84 data-nosnippet>84</a>    /// * `iceberg_schema` - The Iceberg schema reference
-<a href=#85 id=85 data-nosnippet>85</a>    /// * `partition_spec` - The partition specification reference
-<a href=#86 id=86 data-nosnippet>86</a>    ///
-<a href=#87 id=87 data-nosnippet>87</a>    /// # Returns
-<a href=#88 id=88 data-nosnippet>88</a>    ///
-<a href=#89 id=89 data-nosnippet>89</a>    /// Returns a new `RecordBatchPartitionSplitter` instance or an error if initialization fails.
-<a href=#90 id=90 data-nosnippet>90</a>    </span><span class="kw">pub fn </span>try_new_with_computed_values(
-<a href=#91 id=91 data-nosnippet>91</a>        iceberg_schema: SchemaRef,
-<a href=#92 id=92 data-nosnippet>92</a>        partition_spec: PartitionSpecRef,
-<a href=#93 id=93 data-nosnippet>93</a>    ) -&gt; <span class="prelude-ty">Result</span>&lt;<span class="self">Self</span>&gt; {
-<a href=#94 id=94 data-nosnippet>94</a>        <span class="kw">let </span>calculator = PartitionValueCalculator::try_new(<span class="kw-2">&amp;</span>partition_spec, <span class="kw-2">&amp;</span>iceberg_schema)<span class="question-mark">?</span>;
-<a href=#95 id=95 data-nosnippet>95</a>        <span class="self">Self</span>::try_new(iceberg_schema, partition_spec, <span class="prelude-val">Some</span>(calculator))
-<a href=#96 id=96 data-nosnippet>96</a>    }
-<a href=#97 id=97 data-nosnippet>97</a>
-<a href=#98 id=98 data-nosnippet>98</a>    <span class="doccomment">/// Create a new RecordBatchPartitionSplitter expecting pre-computed partition values.
-<a href=#99 id=99 data-nosnippet>99</a>    ///
-<a href=#100 id=100 data-nosnippet>100</a>    /// This is a convenience method that initializes the splitter to expect a `_partition`
-<a href=#101 id=101 data-nosnippet>101</a>    /// column in the input batches.
-<a href=#102 id=102 data-nosnippet>102</a>    ///
-<a href=#103 id=103 data-nosnippet>103</a>    /// # Arguments
-<a href=#104 id=104 data-nosnippet>104</a>    ///
-<a href=#105 id=105 data-nosnippet>105</a>    /// * `iceberg_schema` - The Iceberg schema reference
-<a href=#106 id=106 data-nosnippet>106</a>    /// * `partition_spec` - The partition specification reference
-<a href=#107 id=107 data-nosnippet>107</a>    ///
-<a href=#108 id=108 data-nosnippet>108</a>    /// # Returns
-<a href=#109 id=109 data-nosnippet>109</a>    ///
-<a href=#110 id=110 data-nosnippet>110</a>    /// Returns a new `RecordBatchPartitionSplitter` instance or an error if initialization fails.
-<a href=#111 id=111 data-nosnippet>111</a>    </span><span class="kw">pub fn </span>try_new_with_precomputed_values(
-<a href=#112 id=112 data-nosnippet>112</a>        iceberg_schema: SchemaRef,
-<a href=#113 id=113 data-nosnippet>113</a>        partition_spec: PartitionSpecRef,
-<a href=#114 id=114 data-nosnippet>114</a>    ) -&gt; <span class="prelude-ty">Result</span>&lt;<span class="self">Self</span>&gt; {
-<a href=#115 id=115 data-nosnippet>115</a>        <span class="self">Self</span>::try_new(iceberg_schema, partition_spec, <span class="prelude-val">None</span>)
-<a href=#116 id=116 data-nosnippet>116</a>    }
-<a href=#117 id=117 data-nosnippet>117</a>
-<a href=#118 id=118 data-nosnippet>118</a>    <span class="doccomment">/// Split the record batch into multiple record batches based on the partition spec.
-<a href=#119 id=119 data-nosnippet>119</a>    </span><span class="kw">pub fn </span>split(<span class="kw-2">&amp;</span><span class="self">self</span>, batch: <span class="kw-2">&amp;</span>RecordBatch) -&gt; <span class="prelude-ty">Result</span>&lt;Vec&lt;(PartitionKey, RecordBatch)&gt;&gt; {
-<a href=#120 id=120 data-nosnippet>120</a>        <span class="kw">let </span>partition_structs = <span class="kw">if let </span><span class="prelude-val">Some</span>(calculator) = <span class="kw-2">&amp;</span><span class="self">self</span>.calculator {
-<a href=#121 id=121 data-nosnippet>121</a>            <span class="comment">// Compute partition values from source columns using calculator
-<a href=#122 id=122 data-nosnippet>122</a>            </span><span class="kw">let </span>partition_array = calculator.calculate(batch)<span class="question-mark">?</span>;
-<a href=#123 id=123 data-nosnippet>123</a>            <span class="kw">let </span>struct_array = arrow_struct_to_literal(<span class="kw-2">&amp;</span>partition_array, <span class="kw-2">&amp;</span><span class="self">self</span>.partition_type)<span class="question-mark">?</span>;
-<a href=#124 id=124 data-nosnippet>124</a>
-<a href=#125 id=125 data-nosnippet>125</a>            struct_array
-<a href=#126 id=126 data-nosnippet>126</a>                .into_iter()
-<a href=#127 id=127 data-nosnippet>127</a>                .map(|s| {
-<a href=#128 id=128 data-nosnippet>128</a>                    <span class="kw">if let </span><span class="prelude-val">Some</span>(Literal::Struct(s)) = s {
-<a href=#129 id=129 data-nosnippet>129</a>                        <span class="prelude-val">Ok</span>(s)
-<a href=#130 id=130 data-nosnippet>130</a>                    } <span class="kw">else </span>{
-<a href=#131 id=131 data-nosnippet>131</a>                        <span class="prelude-val">Err</span>(Error::new(
-<a href=#132 id=132 data-nosnippet>132</a>                            ErrorKind::DataInvalid,
-<a href=#133 id=133 data-nosnippet>133</a>                            <span class="string">"Partition value is not a struct literal or is null"</span>,
-<a href=#134 id=134 data-nosnippet>134</a>                        ))
-<a href=#135 id=135 data-nosnippet>135</a>                    }
-<a href=#136 id=136 data-nosnippet>136</a>                })
-<a href=#137 id=137 data-nosnippet>137</a>                .collect::&lt;<span class="prelude-ty">Result</span>&lt;Vec&lt;<span class="kw">_</span>&gt;&gt;&gt;()<span class="question-mark">?
-<a href=#138 id=138 data-nosnippet>138</a>        </span>} <span class="kw">else </span>{
-<a href=#139 id=139 data-nosnippet>139</a>            <span class="comment">// Extract partition values from pre-computed partition column
-<a href=#140 id=140 data-nosnippet>140</a>            </span><span class="kw">let </span>partition_column = batch
-<a href=#141 id=141 data-nosnippet>141</a>                .column_by_name(PROJECTED_PARTITION_VALUE_COLUMN)
-<a href=#142 id=142 data-nosnippet>142</a>                .ok_or_else(|| {
-<a href=#143 id=143 data-nosnippet>143</a>                    Error::new(
-<a href=#144 id=144 data-nosnippet>144</a>                        ErrorKind::DataInvalid,
-<a href=#145 id=145 data-nosnippet>145</a>                        <span class="macro">format!</span>(
-<a href=#146 id=146 data-nosnippet>146</a>                            <span class="string">"Partition column '{PROJECTED_PARTITION_VALUE_COLUMN}' not found in batch"
-<a href=#147 id=147 data-nosnippet>147</a>                        </span>),
-<a href=#148 id=148 data-nosnippet>148</a>                    )
-<a href=#149 id=149 data-nosnippet>149</a>                })<span class="question-mark">?</span>;
-<a href=#150 id=150 data-nosnippet>150</a>
-<a href=#151 id=151 data-nosnippet>151</a>            <span class="kw">let </span>partition_struct_array = partition_column
-<a href=#152 id=152 data-nosnippet>152</a>                .as_any()
-<a href=#153 id=153 data-nosnippet>153</a>                .downcast_ref::&lt;StructArray&gt;()
-<a href=#154 id=154 data-nosnippet>154</a>                .ok_or_else(|| {
-<a href=#155 id=155 data-nosnippet>155</a>                    Error::new(
-<a href=#156 id=156 data-nosnippet>156</a>                        ErrorKind::DataInvalid,
-<a href=#157 id=157 data-nosnippet>157</a>                        <span class="string">"Partition column is not a StructArray"</span>,
-<a href=#158 id=158 data-nosnippet>158</a>                    )
-<a href=#159 id=159 data-nosnippet>159</a>                })<span class="question-mark">?</span>;
-<a href=#160 id=160 data-nosnippet>160</a>
-<a href=#161 id=161 data-nosnippet>161</a>            <span class="kw">let </span>arrow_struct_array = Arc::new(partition_struct_array.clone()) <span class="kw">as </span>ArrayRef;
-<a href=#162 id=162 data-nosnippet>162</a>            <span class="kw">let </span>struct_array = arrow_struct_to_literal(<span class="kw-2">&amp;</span>arrow_struct_array, <span class="kw-2">&amp;</span><span class="self">self</span>.partition_type)<span class="question-mark">?</span>;
-<a href=#163 id=163 data-nosnippet>163</a>
-<a href=#164 id=164 data-nosnippet>164</a>            struct_array
-<a href=#165 id=165 data-nosnippet>165</a>                .into_iter()
-<a href=#166 id=166 data-nosnippet>166</a>                .map(|s| {
-<a href=#167 id=167 data-nosnippet>167</a>                    <span class="kw">if let </span><span class="prelude-val">Some</span>(Literal::Struct(s)) = s {
-<a href=#168 id=168 data-nosnippet>168</a>                        <span class="prelude-val">Ok</span>(s)
-<a href=#169 id=169 data-nosnippet>169</a>                    } <span class="kw">else </span>{
-<a href=#170 id=170 data-nosnippet>170</a>                        <span class="prelude-val">Err</span>(Error::new(
-<a href=#171 id=171 data-nosnippet>171</a>                            ErrorKind::DataInvalid,
-<a href=#172 id=172 data-nosnippet>172</a>                            <span class="string">"Partition value is not a struct literal or is null"</span>,
-<a href=#173 id=173 data-nosnippet>173</a>                        ))
-<a href=#174 id=174 data-nosnippet>174</a>                    }
-<a href=#175 id=175 data-nosnippet>175</a>                })
-<a href=#176 id=176 data-nosnippet>176</a>                .collect::&lt;<span class="prelude-ty">Result</span>&lt;Vec&lt;<span class="kw">_</span>&gt;&gt;&gt;()<span class="question-mark">?
-<a href=#177 id=177 data-nosnippet>177</a>        </span>};
-<a href=#178 id=178 data-nosnippet>178</a>
-<a href=#179 id=179 data-nosnippet>179</a>        <span class="comment">// Group the batch by row value.
-<a href=#180 id=180 data-nosnippet>180</a>        </span><span class="kw">let </span><span class="kw-2">mut </span>group_ids = HashMap::new();
-<a href=#181 id=181 data-nosnippet>181</a>        partition_structs
-<a href=#182 id=182 data-nosnippet>182</a>            .into_iter()
-<a href=#183 id=183 data-nosnippet>183</a>            .enumerate()
-<a href=#184 id=184 data-nosnippet>184</a>            .for_each(|(row_id, row)| {
-<a href=#185 id=185 data-nosnippet>185</a>                group_ids.entry(row).or_insert(<span class="macro">vec!</span>[]).push(row_id);
-<a href=#186 id=186 data-nosnippet>186</a>            });
-<a href=#187 id=187 data-nosnippet>187</a>
-<a href=#188 id=188 data-nosnippet>188</a>        <span class="comment">// Partition the batch with same partition partition_values
-<a href=#189 id=189 data-nosnippet>189</a>        </span><span class="kw">let </span><span class="kw-2">mut </span>partition_batches = Vec::with_capacity(group_ids.len());
-<a href=#190 id=190 data-nosnippet>190</a>        <span class="kw">for </span>(row, row_ids) <span class="kw">in </span>group_ids.into_iter() {
-<a href=#191 id=191 data-nosnippet>191</a>            <span class="comment">// generate the bool filter array from column_ids
-<a href=#192 id=192 data-nosnippet>192</a>            </span><span class="kw">let </span>filter_array: BooleanArray = {
-<a href=#193 id=193 data-nosnippet>193</a>                <span class="kw">let </span><span class="kw-2">mut </span>filter = <span class="macro">vec!</span>[<span class="bool-val">false</span>; batch.num_rows()];
-<a href=#194 id=194 data-nosnippet>194</a>                row_ids.into_iter().for_each(|row_id| {
-<a href=#195 id=195 data-nosnippet>195</a>                    filter[row_id] = <span class="bool-val">true</span>;
-<a href=#196 id=196 data-nosnippet>196</a>                });
-<a href=#197 id=197 data-nosnippet>197</a>                filter.into()
-<a href=#198 id=198 data-nosnippet>198</a>            };
-<a href=#199 id=199 data-nosnippet>199</a>
-<a href=#200 id=200 data-nosnippet>200</a>            <span class="comment">// Create PartitionKey from the partition struct
-<a href=#201 id=201 data-nosnippet>201</a>            </span><span class="kw">let </span>partition_key = PartitionKey::new(
-<a href=#202 id=202 data-nosnippet>202</a>                <span class="self">self</span>.partition_spec.as_ref().clone(),
-<a href=#203 id=203 data-nosnippet>203</a>                <span class="self">self</span>.schema.clone(),
-<a href=#204 id=204 data-nosnippet>204</a>                row,
-<a href=#205 id=205 data-nosnippet>205</a>            );
-<a href=#206 id=206 data-nosnippet>206</a>
-<a href=#207 id=207 data-nosnippet>207</a>            <span class="comment">// filter the RecordBatch
-<a href=#208 id=208 data-nosnippet>208</a>            </span>partition_batches.push((partition_key, filter_record_batch(batch, <span class="kw-2">&amp;</span>filter_array)<span class="question-mark">?</span>));
-<a href=#209 id=209 data-nosnippet>209</a>        }
-<a href=#210 id=210 data-nosnippet>210</a>
-<a href=#211 id=211 data-nosnippet>211</a>        <span class="prelude-val">Ok</span>(partition_batches)
-<a href=#212 id=212 data-nosnippet>212</a>    }
-<a href=#213 id=213 data-nosnippet>213</a>}
-<a href=#214 id=214 data-nosnippet>214</a>
-<a href=#215 id=215 data-nosnippet>215</a><span class="attr">#[cfg(test)]
-<a href=#216 id=216 data-nosnippet>216</a></span><span class="kw">mod </span>tests {
-<a href=#217 id=217 data-nosnippet>217</a>    <span class="kw">use </span>std::sync::Arc;
-<a href=#218 id=218 data-nosnippet>218</a>
-<a href=#219 id=219 data-nosnippet>219</a>    <span class="kw">use </span>arrow_array::{Int32Array, RecordBatch, StringArray};
-<a href=#220 id=220 data-nosnippet>220</a>    <span class="kw">use </span>arrow_schema::DataType;
-<a href=#221 id=221 data-nosnippet>221</a>    <span class="kw">use </span>parquet::arrow::PARQUET_FIELD_ID_META_KEY;
-<a href=#222 id=222 data-nosnippet>222</a>
-<a href=#223 id=223 data-nosnippet>223</a>    <span class="kw">use </span>super::<span class="kw-2">*</span>;
-<a href=#224 id=224 data-nosnippet>224</a>    <span class="kw">use </span><span class="kw">crate</span>::arrow::schema_to_arrow_schema;
-<a href=#225 id=225 data-nosnippet>225</a>    <span class="kw">use </span><span class="kw">crate</span>::spec::{
-<a href=#226 id=226 data-nosnippet>226</a>        NestedField, PartitionSpecBuilder, PrimitiveLiteral, Schema, Struct, Transform, Type,
-<a href=#227 id=227 data-nosnippet>227</a>        UnboundPartitionField,
-<a href=#228 id=228 data-nosnippet>228</a>    };
-<a href=#229 id=229 data-nosnippet>229</a>
-<a href=#230 id=230 data-nosnippet>230</a>    <span class="attr">#[test]
-<a href=#231 id=231 data-nosnippet>231</a>    </span><span class="kw">fn </span>test_record_batch_partition_split() {
-<a href=#232 id=232 data-nosnippet>232</a>        <span class="kw">let </span>schema = Arc::new(
-<a href=#233 id=233 data-nosnippet>233</a>            Schema::builder()
-<a href=#234 id=234 data-nosnippet>234</a>                .with_fields(<span class="macro">vec!</span>[
-<a href=#235 id=235 data-nosnippet>235</a>                    NestedField::required(
-<a href=#236 id=236 data-nosnippet>236</a>                        <span class="number">1</span>,
-<a href=#237 id=237 data-nosnippet>237</a>                        <span class="string">"id"</span>,
-<a href=#238 id=238 data-nosnippet>238</a>                        Type::Primitive(<span class="kw">crate</span>::spec::PrimitiveType::Int),
-<a href=#239 id=239 data-nosnippet>239</a>                    )
-<a href=#240 id=240 data-nosnippet>240</a>                    .into(),
-<a href=#241 id=241 data-nosnippet>241</a>                    NestedField::required(
-<a href=#242 id=242 data-nosnippet>242</a>                        <span class="number">2</span>,
-<a href=#243 id=243 data-nosnippet>243</a>                        <span class="string">"name"</span>,
-<a href=#244 id=244 data-nosnippet>244</a>                        Type::Primitive(<span class="kw">crate</span>::spec::PrimitiveType::String),
-<a href=#245 id=245 data-nosnippet>245</a>                    )
-<a href=#246 id=246 data-nosnippet>246</a>                    .into(),
-<a href=#247 id=247 data-nosnippet>247</a>                ])
-<a href=#248 id=248 data-nosnippet>248</a>                .build()
-<a href=#249 id=249 data-nosnippet>249</a>                .unwrap(),
-<a href=#250 id=250 data-nosnippet>250</a>        );
-<a href=#251 id=251 data-nosnippet>251</a>        <span class="kw">let </span>partition_spec = Arc::new(
-<a href=#252 id=252 data-nosnippet>252</a>            PartitionSpecBuilder::new(schema.clone())
-<a href=#253 id=253 data-nosnippet>253</a>                .with_spec_id(<span class="number">1</span>)
-<a href=#254 id=254 data-nosnippet>254</a>                .add_unbound_field(UnboundPartitionField {
-<a href=#255 id=255 data-nosnippet>255</a>                    source_id: <span class="number">1</span>,
-<a href=#256 id=256 data-nosnippet>256</a>                    field_id: <span class="prelude-val">None</span>,
-<a href=#257 id=257 data-nosnippet>257</a>                    name: <span class="string">"id_bucket"</span>.to_string(),
-<a href=#258 id=258 data-nosnippet>258</a>                    transform: Transform::Identity,
-<a href=#259 id=259 data-nosnippet>259</a>                })
-<a href=#260 id=260 data-nosnippet>260</a>                .unwrap()
-<a href=#261 id=261 data-nosnippet>261</a>                .build()
-<a href=#262 id=262 data-nosnippet>262</a>                .unwrap(),
-<a href=#263 id=263 data-nosnippet>263</a>        );
-<a href=#264 id=264 data-nosnippet>264</a>        <span class="kw">let </span>partition_splitter = RecordBatchPartitionSplitter::try_new_with_computed_values(
-<a href=#265 id=265 data-nosnippet>265</a>            schema.clone(),
-<a href=#266 id=266 data-nosnippet>266</a>            partition_spec,
-<a href=#267 id=267 data-nosnippet>267</a>        )
-<a href=#268 id=268 data-nosnippet>268</a>        .expect(<span class="string">"Failed to create splitter"</span>);
-<a href=#269 id=269 data-nosnippet>269</a>
-<a href=#270 id=270 data-nosnippet>270</a>        <span class="kw">let </span>arrow_schema = Arc::new(schema_to_arrow_schema(<span class="kw-2">&amp;</span>schema).unwrap());
-<a href=#271 id=271 data-nosnippet>271</a>        <span class="kw">let </span>id_array = Int32Array::from(<span class="macro">vec!</span>[<span class="number">1</span>, <span class="number">2</span>, <span class="number">1</span>, <span class="number">3</span>, <span class="number">2</span>, <span class="number">3</span>, <span class="number">1</span>]);
-<a href=#272 id=272 data-nosnippet>272</a>        <span class="kw">let </span>data_array = StringArray::from(<span class="macro">vec!</span>[<span class="string">"a"</span>, <span class="string">"b"</span>, <span class="string">"c"</span>, <span class="string">"d"</span>, <span class="string">"e"</span>, <span class="string">"f"</span>, <span class="string">"g"</span>]);
-<a href=#273 id=273 data-nosnippet>273</a>        <span class="kw">let </span>batch = RecordBatch::try_new(arrow_schema.clone(), <span class="macro">vec!</span>[
-<a href=#274 id=274 data-nosnippet>274</a>            Arc::new(id_array),
-<a href=#275 id=275 data-nosnippet>275</a>            Arc::new(data_array),
-<a href=#276 id=276 data-nosnippet>276</a>        ])
-<a href=#277 id=277 data-nosnippet>277</a>        .expect(<span class="string">"Failed to create RecordBatch"</span>);
-<a href=#278 id=278 data-nosnippet>278</a>
-<a href=#279 id=279 data-nosnippet>279</a>        <span class="kw">let </span><span class="kw-2">mut </span>partitioned_batches = partition_splitter
-<a href=#280 id=280 data-nosnippet>280</a>            .split(<span class="kw-2">&amp;</span>batch)
-<a href=#281 id=281 data-nosnippet>281</a>            .expect(<span class="string">"Failed to split RecordBatch"</span>);
-<a href=#282 id=282 data-nosnippet>282</a>        partitioned_batches.sort_by_key(|(partition_key, <span class="kw">_</span>)| {
-<a href=#283 id=283 data-nosnippet>283</a>            <span class="kw">if let </span>PrimitiveLiteral::Int(i) = partition_key.data().fields()[<span class="number">0</span>]
-<a href=#284 id=284 data-nosnippet>284</a>                .as_ref()
-<a href=#285 id=285 data-nosnippet>285</a>                .unwrap()
-<a href=#286 id=286 data-nosnippet>286</a>                .as_primitive_literal()
+<a href=#22 id=22 data-nosnippet>22</a><span class="kw">use </span>arrow_buffer::BooleanBufferBuilder;
+<a href=#23 id=23 data-nosnippet>23</a><span class="kw">use </span>arrow_select::filter::filter_record_batch;
+<a href=#24 id=24 data-nosnippet>24</a>
+<a href=#25 id=25 data-nosnippet>25</a><span class="kw">use </span><span class="kw">super</span>::arrow_struct_to_literal;
+<a href=#26 id=26 data-nosnippet>26</a><span class="kw">use </span><span class="kw">super</span>::partition_value_calculator::PartitionValueCalculator;
+<a href=#27 id=27 data-nosnippet>27</a><span class="kw">use </span><span class="kw">crate</span>::spec::{Literal, PartitionKey, PartitionSpecRef, SchemaRef, StructType};
+<a href=#28 id=28 data-nosnippet>28</a><span class="kw">use </span>crate::{Error, ErrorKind, <span class="prelude-ty">Result</span>};
+<a href=#29 id=29 data-nosnippet>29</a>
+<a href=#30 id=30 data-nosnippet>30</a><span class="doccomment">/// Column name for the projected partition values struct
+<a href=#31 id=31 data-nosnippet>31</a></span><span class="kw">pub const </span>PROJECTED_PARTITION_VALUE_COLUMN: <span class="kw-2">&amp;</span>str = <span class="string">"_partition"</span>;
+<a href=#32 id=32 data-nosnippet>32</a>
+<a href=#33 id=33 data-nosnippet>33</a><span class="doccomment">/// The splitter used to split the record batch into multiple record batches by the partition spec.
+<a href=#34 id=34 data-nosnippet>34</a>/// 1. It will project and transform the input record batch based on the partition spec, get the partitioned record batch.
+<a href=#35 id=35 data-nosnippet>35</a>/// 2. Split the input record batch into multiple record batches based on the partitioned record batch.
+<a href=#36 id=36 data-nosnippet>36</a>///
+<a href=#37 id=37 data-nosnippet>37</a>/// # Partition Value Modes
+<a href=#38 id=38 data-nosnippet>38</a>///
+<a href=#39 id=39 data-nosnippet>39</a>/// The splitter supports two modes for obtaining partition values:
+<a href=#40 id=40 data-nosnippet>40</a>/// - **Computed mode** (`calculator` is `Some`): Computes partition values from source columns using transforms
+<a href=#41 id=41 data-nosnippet>41</a>/// - **Pre-computed mode** (`calculator` is `None`): Expects a `_partition` column in the input batch
+<a href=#42 id=42 data-nosnippet>42</a></span><span class="kw">pub struct </span>RecordBatchPartitionSplitter {
+<a href=#43 id=43 data-nosnippet>43</a>    schema: SchemaRef,
+<a href=#44 id=44 data-nosnippet>44</a>    partition_spec: PartitionSpecRef,
+<a href=#45 id=45 data-nosnippet>45</a>    calculator: <span class="prelude-ty">Option</span>&lt;PartitionValueCalculator&gt;,
+<a href=#46 id=46 data-nosnippet>46</a>    partition_type: StructType,
+<a href=#47 id=47 data-nosnippet>47</a>}
+<a href=#48 id=48 data-nosnippet>48</a>
+<a href=#49 id=49 data-nosnippet>49</a><span class="kw">impl </span>RecordBatchPartitionSplitter {
+<a href=#50 id=50 data-nosnippet>50</a>    <span class="doccomment">/// Create a new RecordBatchPartitionSplitter.
+<a href=#51 id=51 data-nosnippet>51</a>    ///
+<a href=#52 id=52 data-nosnippet>52</a>    /// # Arguments
+<a href=#53 id=53 data-nosnippet>53</a>    ///
+<a href=#54 id=54 data-nosnippet>54</a>    /// * `iceberg_schema` - The Iceberg schema reference
+<a href=#55 id=55 data-nosnippet>55</a>    /// * `partition_spec` - The partition specification reference
+<a href=#56 id=56 data-nosnippet>56</a>    /// * `calculator` - Optional calculator for computing partition values from source columns.
+<a href=#57 id=57 data-nosnippet>57</a>    ///   - `Some(calculator)`: Compute partition values from source columns using transforms
+<a href=#58 id=58 data-nosnippet>58</a>    ///   - `None`: Expect a pre-computed `_partition` column in the input batch
+<a href=#59 id=59 data-nosnippet>59</a>    ///
+<a href=#60 id=60 data-nosnippet>60</a>    /// # Returns
+<a href=#61 id=61 data-nosnippet>61</a>    ///
+<a href=#62 id=62 data-nosnippet>62</a>    /// Returns a new `RecordBatchPartitionSplitter` instance or an error if initialization fails.
+<a href=#63 id=63 data-nosnippet>63</a>    </span><span class="kw">pub fn </span>try_new(
+<a href=#64 id=64 data-nosnippet>64</a>        iceberg_schema: SchemaRef,
+<a href=#65 id=65 data-nosnippet>65</a>        partition_spec: PartitionSpecRef,
+<a href=#66 id=66 data-nosnippet>66</a>        calculator: <span class="prelude-ty">Option</span>&lt;PartitionValueCalculator&gt;,
+<a href=#67 id=67 data-nosnippet>67</a>    ) -&gt; <span class="prelude-ty">Result</span>&lt;<span class="self">Self</span>&gt; {
+<a href=#68 id=68 data-nosnippet>68</a>        <span class="kw">let </span>partition_type = partition_spec.partition_type(<span class="kw-2">&amp;</span>iceberg_schema)<span class="question-mark">?</span>;
+<a href=#69 id=69 data-nosnippet>69</a>
+<a href=#70 id=70 data-nosnippet>70</a>        <span class="prelude-val">Ok</span>(<span class="self">Self </span>{
+<a href=#71 id=71 data-nosnippet>71</a>            schema: iceberg_schema,
+<a href=#72 id=72 data-nosnippet>72</a>            partition_spec,
+<a href=#73 id=73 data-nosnippet>73</a>            calculator,
+<a href=#74 id=74 data-nosnippet>74</a>            partition_type,
+<a href=#75 id=75 data-nosnippet>75</a>        })
+<a href=#76 id=76 data-nosnippet>76</a>    }
+<a href=#77 id=77 data-nosnippet>77</a>
+<a href=#78 id=78 data-nosnippet>78</a>    <span class="doccomment">/// Create a new RecordBatchPartitionSplitter with computed partition values.
+<a href=#79 id=79 data-nosnippet>79</a>    ///
+<a href=#80 id=80 data-nosnippet>80</a>    /// This is a convenience method that creates a calculator and initializes the splitter
+<a href=#81 id=81 data-nosnippet>81</a>    /// to compute partition values from source columns.
+<a href=#82 id=82 data-nosnippet>82</a>    ///
+<a href=#83 id=83 data-nosnippet>83</a>    /// # Arguments
+<a href=#84 id=84 data-nosnippet>84</a>    ///
+<a href=#85 id=85 data-nosnippet>85</a>    /// * `iceberg_schema` - The Iceberg schema reference
+<a href=#86 id=86 data-nosnippet>86</a>    /// * `partition_spec` - The partition specification reference
+<a href=#87 id=87 data-nosnippet>87</a>    ///
+<a href=#88 id=88 data-nosnippet>88</a>    /// # Returns
+<a href=#89 id=89 data-nosnippet>89</a>    ///
+<a href=#90 id=90 data-nosnippet>90</a>    /// Returns a new `RecordBatchPartitionSplitter` instance or an error if initialization fails.
+<a href=#91 id=91 data-nosnippet>91</a>    </span><span class="kw">pub fn </span>try_new_with_computed_values(
+<a href=#92 id=92 data-nosnippet>92</a>        iceberg_schema: SchemaRef,
+<a href=#93 id=93 data-nosnippet>93</a>        partition_spec: PartitionSpecRef,
+<a href=#94 id=94 data-nosnippet>94</a>    ) -&gt; <span class="prelude-ty">Result</span>&lt;<span class="self">Self</span>&gt; {
+<a href=#95 id=95 data-nosnippet>95</a>        <span class="kw">let </span>calculator = PartitionValueCalculator::try_new(<span class="kw-2">&amp;</span>partition_spec, <span class="kw-2">&amp;</span>iceberg_schema)<span class="question-mark">?</span>;
+<a href=#96 id=96 data-nosnippet>96</a>        <span class="self">Self</span>::try_new(iceberg_schema, partition_spec, <span class="prelude-val">Some</span>(calculator))
+<a href=#97 id=97 data-nosnippet>97</a>    }
+<a href=#98 id=98 data-nosnippet>98</a>
+<a href=#99 id=99 data-nosnippet>99</a>    <span class="doccomment">/// Create a new RecordBatchPartitionSplitter expecting pre-computed partition values.
+<a href=#100 id=100 data-nosnippet>100</a>    ///
+<a href=#101 id=101 data-nosnippet>101</a>    /// This is a convenience method that initializes the splitter to expect a `_partition`
+<a href=#102 id=102 data-nosnippet>102</a>    /// column in the input batches.
+<a href=#103 id=103 data-nosnippet>103</a>    ///
+<a href=#104 id=104 data-nosnippet>104</a>    /// # Arguments
+<a href=#105 id=105 data-nosnippet>105</a>    ///
+<a href=#106 id=106 data-nosnippet>106</a>    /// * `iceberg_schema` - The Iceberg schema reference
+<a href=#107 id=107 data-nosnippet>107</a>    /// * `partition_spec` - The partition specification reference
+<a href=#108 id=108 data-nosnippet>108</a>    ///
+<a href=#109 id=109 data-nosnippet>109</a>    /// # Returns
+<a href=#110 id=110 data-nosnippet>110</a>    ///
+<a href=#111 id=111 data-nosnippet>111</a>    /// Returns a new `RecordBatchPartitionSplitter` instance or an error if initialization fails.
+<a href=#112 id=112 data-nosnippet>112</a>    </span><span class="kw">pub fn </span>try_new_with_precomputed_values(
+<a href=#113 id=113 data-nosnippet>113</a>        iceberg_schema: SchemaRef,
+<a href=#114 id=114 data-nosnippet>114</a>        partition_spec: PartitionSpecRef,
+<a href=#115 id=115 data-nosnippet>115</a>    ) -&gt; <span class="prelude-ty">Result</span>&lt;<span class="self">Self</span>&gt; {
+<a href=#116 id=116 data-nosnippet>116</a>        <span class="self">Self</span>::try_new(iceberg_schema, partition_spec, <span class="prelude-val">None</span>)
+<a href=#117 id=117 data-nosnippet>117</a>    }
+<a href=#118 id=118 data-nosnippet>118</a>
+<a href=#119 id=119 data-nosnippet>119</a>    <span class="doccomment">/// Split the record batch into multiple record batches based on the partition spec.
+<a href=#120 id=120 data-nosnippet>120</a>    </span><span class="kw">pub fn </span>split(<span class="kw-2">&amp;</span><span class="self">self</span>, batch: <span class="kw-2">&amp;</span>RecordBatch) -&gt; <span class="prelude-ty">Result</span>&lt;Vec&lt;(PartitionKey, RecordBatch)&gt;&gt; {
+<a href=#121 id=121 data-nosnippet>121</a>        <span class="kw">let </span>partition_structs = <span class="kw">if let </span><span class="prelude-val">Some</span>(calculator) = <span class="kw-2">&amp;</span><span class="self">self</span>.calculator {
+<a href=#122 id=122 data-nosnippet>122</a>            <span class="comment">// Compute partition values from source columns using calculator
+<a href=#123 id=123 data-nosnippet>123</a>            </span><span class="kw">let </span>partition_array = calculator.calculate(batch)<span class="question-mark">?</span>;
+<a href=#124 id=124 data-nosnippet>124</a>            <span class="kw">let </span>struct_array = arrow_struct_to_literal(<span class="kw-2">&amp;</span>partition_array, <span class="kw-2">&amp;</span><span class="self">self</span>.partition_type)<span class="question-mark">?</span>;
+<a href=#125 id=125 data-nosnippet>125</a>
+<a href=#126 id=126 data-nosnippet>126</a>            struct_array
+<a href=#127 id=127 data-nosnippet>127</a>                .into_iter()
+<a href=#128 id=128 data-nosnippet>128</a>                .map(|s| {
+<a href=#129 id=129 data-nosnippet>129</a>                    <span class="kw">if let </span><span class="prelude-val">Some</span>(Literal::Struct(s)) = s {
+<a href=#130 id=130 data-nosnippet>130</a>                        <span class="prelude-val">Ok</span>(s)
+<a href=#131 id=131 data-nosnippet>131</a>                    } <span class="kw">else </span>{
+<a href=#132 id=132 data-nosnippet>132</a>                        <span class="prelude-val">Err</span>(Error::new(
+<a href=#133 id=133 data-nosnippet>133</a>                            ErrorKind::DataInvalid,
+<a href=#134 id=134 data-nosnippet>134</a>                            <span class="string">"Partition value is not a struct literal or is null"</span>,
+<a href=#135 id=135 data-nosnippet>135</a>                        ))
+<a href=#136 id=136 data-nosnippet>136</a>                    }
+<a href=#137 id=137 data-nosnippet>137</a>                })
+<a href=#138 id=138 data-nosnippet>138</a>                .collect::&lt;<span class="prelude-ty">Result</span>&lt;Vec&lt;<span class="kw">_</span>&gt;&gt;&gt;()<span class="question-mark">?
+<a href=#139 id=139 data-nosnippet>139</a>        </span>} <span class="kw">else </span>{
+<a href=#140 id=140 data-nosnippet>140</a>            <span class="comment">// Extract partition values from pre-computed partition column
+<a href=#141 id=141 data-nosnippet>141</a>            </span><span class="kw">let </span>partition_column = batch
+<a href=#142 id=142 data-nosnippet>142</a>                .column_by_name(PROJECTED_PARTITION_VALUE_COLUMN)
+<a href=#143 id=143 data-nosnippet>143</a>                .ok_or_else(|| {
+<a href=#144 id=144 data-nosnippet>144</a>                    Error::new(
+<a href=#145 id=145 data-nosnippet>145</a>                        ErrorKind::DataInvalid,
+<a href=#146 id=146 data-nosnippet>146</a>                        <span class="macro">format!</span>(
+<a href=#147 id=147 data-nosnippet>147</a>                            <span class="string">"Partition column '{PROJECTED_PARTITION_VALUE_COLUMN}' not found in batch"
+<a href=#148 id=148 data-nosnippet>148</a>                        </span>),
+<a href=#149 id=149 data-nosnippet>149</a>                    )
+<a href=#150 id=150 data-nosnippet>150</a>                })<span class="question-mark">?</span>;
+<a href=#151 id=151 data-nosnippet>151</a>
+<a href=#152 id=152 data-nosnippet>152</a>            <span class="kw">let </span>partition_struct_array = partition_column
+<a href=#153 id=153 data-nosnippet>153</a>                .as_any()
+<a href=#154 id=154 data-nosnippet>154</a>                .downcast_ref::&lt;StructArray&gt;()
+<a href=#155 id=155 data-nosnippet>155</a>                .ok_or_else(|| {
+<a href=#156 id=156 data-nosnippet>156</a>                    Error::new(
+<a href=#157 id=157 data-nosnippet>157</a>                        ErrorKind::DataInvalid,
+<a href=#158 id=158 data-nosnippet>158</a>                        <span class="string">"Partition column is not a StructArray"</span>,
+<a href=#159 id=159 data-nosnippet>159</a>                    )
+<a href=#160 id=160 data-nosnippet>160</a>                })<span class="question-mark">?</span>;
+<a href=#161 id=161 data-nosnippet>161</a>
+<a href=#162 id=162 data-nosnippet>162</a>            <span class="kw">let </span>arrow_struct_array = Arc::new(partition_struct_array.clone()) <span class="kw">as </span>ArrayRef;
+<a href=#163 id=163 data-nosnippet>163</a>            <span class="kw">let </span>struct_array = arrow_struct_to_literal(<span class="kw-2">&amp;</span>arrow_struct_array, <span class="kw-2">&amp;</span><span class="self">self</span>.partition_type)<span class="question-mark">?</span>;
+<a href=#164 id=164 data-nosnippet>164</a>
+<a href=#165 id=165 data-nosnippet>165</a>            struct_array
+<a href=#166 id=166 data-nosnippet>166</a>                .into_iter()
+<a href=#167 id=167 data-nosnippet>167</a>                .map(|s| {
+<a href=#168 id=168 data-nosnippet>168</a>                    <span class="kw">if let </span><span class="prelude-val">Some</span>(Literal::Struct(s)) = s {
+<a href=#169 id=169 data-nosnippet>169</a>                        <span class="prelude-val">Ok</span>(s)
+<a href=#170 id=170 data-nosnippet>170</a>                    } <span class="kw">else </span>{
+<a href=#171 id=171 data-nosnippet>171</a>                        <span class="prelude-val">Err</span>(Error::new(
+<a href=#172 id=172 data-nosnippet>172</a>                            ErrorKind::DataInvalid,
+<a href=#173 id=173 data-nosnippet>173</a>                            <span class="string">"Partition value is not a struct literal or is null"</span>,
+<a href=#174 id=174 data-nosnippet>174</a>                        ))
+<a href=#175 id=175 data-nosnippet>175</a>                    }
+<a href=#176 id=176 data-nosnippet>176</a>                })
+<a href=#177 id=177 data-nosnippet>177</a>                .collect::&lt;<span class="prelude-ty">Result</span>&lt;Vec&lt;<span class="kw">_</span>&gt;&gt;&gt;()<span class="question-mark">?
+<a href=#178 id=178 data-nosnippet>178</a>        </span>};
+<a href=#179 id=179 data-nosnippet>179</a>
+<a href=#180 id=180 data-nosnippet>180</a>        <span class="comment">// Group the batch by row value.
+<a href=#181 id=181 data-nosnippet>181</a>        </span><span class="kw">let </span><span class="kw-2">mut </span>group_ids = HashMap::new();
+<a href=#182 id=182 data-nosnippet>182</a>        partition_structs
+<a href=#183 id=183 data-nosnippet>183</a>            .into_iter()
+<a href=#184 id=184 data-nosnippet>184</a>            .enumerate()
+<a href=#185 id=185 data-nosnippet>185</a>            .for_each(|(row_id, row)| {
+<a href=#186 id=186 data-nosnippet>186</a>                group_ids.entry(row).or_insert(<span class="macro">vec!</span>[]).push(row_id);
+<a href=#187 id=187 data-nosnippet>187</a>            });
+<a href=#188 id=188 data-nosnippet>188</a>
+<a href=#189 id=189 data-nosnippet>189</a>        <span class="comment">// Partition the batch with same partition partition_values
+<a href=#190 id=190 data-nosnippet>190</a>        </span><span class="kw">let </span><span class="kw-2">mut </span>partition_batches = Vec::with_capacity(group_ids.len());
+<a href=#191 id=191 data-nosnippet>191</a>        <span class="kw">for </span>(row, row_ids) <span class="kw">in </span>group_ids.into_iter() {
+<a href=#192 id=192 data-nosnippet>192</a>            <span class="comment">// generate the bool filter array from column_ids
+<a href=#193 id=193 data-nosnippet>193</a>            </span><span class="kw">let </span>filter_array: BooleanArray = {
+<a href=#194 id=194 data-nosnippet>194</a>                <span class="kw">let </span><span class="kw-2">mut </span>builder = BooleanBufferBuilder::new(batch.num_rows());
+<a href=#195 id=195 data-nosnippet>195</a>                builder.append_n(batch.num_rows(), <span class="bool-val">false</span>);
+<a href=#196 id=196 data-nosnippet>196</a>                <span class="kw">for </span>row_id <span class="kw">in </span>row_ids {
+<a href=#197 id=197 data-nosnippet>197</a>                    builder.set_bit(row_id, <span class="bool-val">true</span>);
+<a href=#198 id=198 data-nosnippet>198</a>                }
+<a href=#199 id=199 data-nosnippet>199</a>                BooleanArray::new(builder.finish(), <span class="prelude-val">None</span>)
+<a href=#200 id=200 data-nosnippet>200</a>            };
+<a href=#201 id=201 data-nosnippet>201</a>
+<a href=#202 id=202 data-nosnippet>202</a>            <span class="comment">// Create PartitionKey from the partition struct
+<a href=#203 id=203 data-nosnippet>203</a>            </span><span class="kw">let </span>partition_key = PartitionKey::new(
+<a href=#204 id=204 data-nosnippet>204</a>                <span class="self">self</span>.partition_spec.as_ref().clone(),
+<a href=#205 id=205 data-nosnippet>205</a>                <span class="self">self</span>.schema.clone(),
+<a href=#206 id=206 data-nosnippet>206</a>                row,
+<a href=#207 id=207 data-nosnippet>207</a>            );
+<a href=#208 id=208 data-nosnippet>208</a>
+<a href=#209 id=209 data-nosnippet>209</a>            <span class="comment">// filter the RecordBatch
+<a href=#210 id=210 data-nosnippet>210</a>            </span>partition_batches.push((partition_key, filter_record_batch(batch, <span class="kw-2">&amp;</span>filter_array)<span class="question-mark">?</span>));
+<a href=#211 id=211 data-nosnippet>211</a>        }
+<a href=#212 id=212 data-nosnippet>212</a>
+<a href=#213 id=213 data-nosnippet>213</a>        <span class="prelude-val">Ok</span>(partition_batches)
+<a href=#214 id=214 data-nosnippet>214</a>    }
+<a href=#215 id=215 data-nosnippet>215</a>}
+<a href=#216 id=216 data-nosnippet>216</a>
+<a href=#217 id=217 data-nosnippet>217</a><span class="attr">#[cfg(test)]
+<a href=#218 id=218 data-nosnippet>218</a></span><span class="kw">mod </span>tests {
+<a href=#219 id=219 data-nosnippet>219</a>    <span class="kw">use </span>std::sync::Arc;
+<a href=#220 id=220 data-nosnippet>220</a>
+<a href=#221 id=221 data-nosnippet>221</a>    <span class="kw">use </span>arrow_array::{Int32Array, RecordBatch, StringArray};
+<a href=#222 id=222 data-nosnippet>222</a>    <span class="kw">use </span>arrow_schema::DataType;
+<a href=#223 id=223 data-nosnippet>223</a>    <span class="kw">use </span>parquet::arrow::PARQUET_FIELD_ID_META_KEY;
+<a href=#224 id=224 data-nosnippet>224</a>
+<a href=#225 id=225 data-nosnippet>225</a>    <span class="kw">use </span>super::<span class="kw-2">*</span>;
+<a href=#226 id=226 data-nosnippet>226</a>    <span class="kw">use </span><span class="kw">crate</span>::arrow::schema_to_arrow_schema;
+<a href=#227 id=227 data-nosnippet>227</a>    <span class="kw">use </span><span class="kw">crate</span>::spec::{
+<a href=#228 id=228 data-nosnippet>228</a>        NestedField, PartitionSpecBuilder, PrimitiveLiteral, Schema, Struct, Transform, Type,
+<a href=#229 id=229 data-nosnippet>229</a>        UnboundPartitionField,
+<a href=#230 id=230 data-nosnippet>230</a>    };
+<a href=#231 id=231 data-nosnippet>231</a>
+<a href=#232 id=232 data-nosnippet>232</a>    <span class="attr">#[test]
+<a href=#233 id=233 data-nosnippet>233</a>    </span><span class="kw">fn </span>test_record_batch_partition_split() {
+<a href=#234 id=234 data-nosnippet>234</a>        <span class="kw">let </span>schema = Arc::new(
+<a href=#235 id=235 data-nosnippet>235</a>            Schema::builder()
+<a href=#236 id=236 data-nosnippet>236</a>                .with_fields(<span class="macro">vec!</span>[
+<a href=#237 id=237 data-nosnippet>237</a>                    NestedField::required(
+<a href=#238 id=238 data-nosnippet>238</a>                        <span class="number">1</span>,
+<a href=#239 id=239 data-nosnippet>239</a>                        <span class="string">"id"</span>,
+<a href=#240 id=240 data-nosnippet>240</a>                        Type::Primitive(<span class="kw">crate</span>::spec::PrimitiveType::Int),
+<a href=#241 id=241 data-nosnippet>241</a>                    )
+<a href=#242 id=242 data-nosnippet>242</a>                    .into(),
+<a href=#243 id=243 data-nosnippet>243</a>                    NestedField::required(
+<a href=#244 id=244 data-nosnippet>244</a>                        <span class="number">2</span>,
+<a href=#245 id=245 data-nosnippet>245</a>                        <span class="string">"name"</span>,
+<a href=#246 id=246 data-nosnippet>246</a>                        Type::Primitive(<span class="kw">crate</span>::spec::PrimitiveType::String),
+<a href=#247 id=247 data-nosnippet>247</a>                    )
+<a href=#248 id=248 data-nosnippet>248</a>                    .into(),
+<a href=#249 id=249 data-nosnippet>249</a>                ])
+<a href=#250 id=250 data-nosnippet>250</a>                .build()
+<a href=#251 id=251 data-nosnippet>251</a>                .unwrap(),
+<a href=#252 id=252 data-nosnippet>252</a>        );
+<a href=#253 id=253 data-nosnippet>253</a>        <span class="kw">let </span>partition_spec = Arc::new(
+<a href=#254 id=254 data-nosnippet>254</a>            PartitionSpecBuilder::new(schema.clone())
+<a href=#255 id=255 data-nosnippet>255</a>                .with_spec_id(<span class="number">1</span>)
+<a href=#256 id=256 data-nosnippet>256</a>                .add_unbound_field(UnboundPartitionField {
+<a href=#257 id=257 data-nosnippet>257</a>                    source_id: <span class="number">1</span>,
+<a href=#258 id=258 data-nosnippet>258</a>                    field_id: <span class="prelude-val">None</span>,
+<a href=#259 id=259 data-nosnippet>259</a>                    name: <span class="string">"id_bucket"</span>.to_string(),
+<a href=#260 id=260 data-nosnippet>260</a>                    transform: Transform::Identity,
+<a href=#261 id=261 data-nosnippet>261</a>                })
+<a href=#262 id=262 data-nosnippet>262</a>                .unwrap()
+<a href=#263 id=263 data-nosnippet>263</a>                .build()
+<a href=#264 id=264 data-nosnippet>264</a>                .unwrap(),
+<a href=#265 id=265 data-nosnippet>265</a>        );
+<a href=#266 id=266 data-nosnippet>266</a>        <span class="kw">let </span>partition_splitter = RecordBatchPartitionSplitter::try_new_with_computed_values(
+<a href=#267 id=267 data-nosnippet>267</a>            schema.clone(),
+<a href=#268 id=268 data-nosnippet>268</a>            partition_spec,
+<a href=#269 id=269 data-nosnippet>269</a>        )
+<a href=#270 id=270 data-nosnippet>270</a>        .expect(<span class="string">"Failed to create splitter"</span>);
+<a href=#271 id=271 data-nosnippet>271</a>
+<a href=#272 id=272 data-nosnippet>272</a>        <span class="kw">let </span>arrow_schema = Arc::new(schema_to_arrow_schema(<span class="kw-2">&amp;</span>schema).unwrap());
+<a href=#273 id=273 data-nosnippet>273</a>        <span class="kw">let </span>id_array = Int32Array::from(<span class="macro">vec!</span>[<span class="number">1</span>, <span class="number">2</span>, <span class="number">1</span>, <span class="number">3</span>, <span class="number">2</span>, <span class="number">3</span>, <span class="number">1</span>]);
+<a href=#274 id=274 data-nosnippet>274</a>        <span class="kw">let </span>data_array = StringArray::from(<span class="macro">vec!</span>[<span class="string">"a"</span>, <span class="string">"b"</span>, <span class="string">"c"</span>, <span class="string">"d"</span>, <span class="string">"e"</span>, <span class="string">"f"</span>, <span class="string">"g"</span>]);
+<a href=#275 id=275 data-nosnippet>275</a>        <span class="kw">let </span>batch = RecordBatch::try_new(arrow_schema.clone(), <span class="macro">vec!</span>[
+<a href=#276 id=276 data-nosnippet>276</a>            Arc::new(id_array),
+<a href=#277 id=277 data-nosnippet>277</a>            Arc::new(data_array),
+<a href=#278 id=278 data-nosnippet>278</a>        ])
+<a href=#279 id=279 data-nosnippet>279</a>        .expect(<span class="string">"Failed to create RecordBatch"</span>);
+<a href=#280 id=280 data-nosnippet>280</a>
+<a href=#281 id=281 data-nosnippet>281</a>        <span class="kw">let </span><span class="kw-2">mut </span>partitioned_batches = partition_splitter
+<a href=#282 id=282 data-nosnippet>282</a>            .split(<span class="kw-2">&amp;</span>batch)
+<a href=#283 id=283 data-nosnippet>283</a>            .expect(<span class="string">"Failed to split RecordBatch"</span>);
+<a href=#284 id=284 data-nosnippet>284</a>        partitioned_batches.sort_by_key(|(partition_key, <span class="kw">_</span>)| {
+<a href=#285 id=285 data-nosnippet>285</a>            <span class="kw">if let </span>PrimitiveLiteral::Int(i) = partition_key.data().fields()[<span class="number">0</span>]
+<a href=#286 id=286 data-nosnippet>286</a>                .as_ref()
 <a href=#287 id=287 data-nosnippet>287</a>                .unwrap()
-<a href=#288 id=288 data-nosnippet>288</a>            {
-<a href=#289 id=289 data-nosnippet>289</a>                i
-<a href=#290 id=290 data-nosnippet>290</a>            } <span class="kw">else </span>{
-<a href=#291 id=291 data-nosnippet>291</a>                <span class="macro">panic!</span>(<span class="string">"The partition value is not a int"</span>);
-<a href=#292 id=292 data-nosnippet>292</a>            }
-<a href=#293 id=293 data-nosnippet>293</a>        });
-<a href=#294 id=294 data-nosnippet>294</a>        <span class="macro">assert_eq!</span>(partitioned_batches.len(), <span class="number">3</span>);
-<a href=#295 id=295 data-nosnippet>295</a>        {
-<a href=#296 id=296 data-nosnippet>296</a>            <span class="comment">// check the first partition
-<a href=#297 id=297 data-nosnippet>297</a>            </span><span class="kw">let </span>expected_id_array = Int32Array::from(<span class="macro">vec!</span>[<span class="number">1</span>, <span class="number">1</span>, <span class="number">1</span>]);
-<a href=#298 id=298 data-nosnippet>298</a>            <span class="kw">let </span>expected_data_array = StringArray::from(<span class="macro">vec!</span>[<span class="string">"a"</span>, <span class="string">"c"</span>, <span class="string">"g"</span>]);
-<a href=#299 id=299 data-nosnippet>299</a>            <span class="kw">let </span>expected_batch = RecordBatch::try_new(arrow_schema.clone(), <span class="macro">vec!</span>[
-<a href=#300 id=300 data-nosnippet>300</a>                Arc::new(expected_id_array),
-<a href=#301 id=301 data-nosnippet>301</a>                Arc::new(expected_data_array),
-<a href=#302 id=302 data-nosnippet>302</a>            ])
-<a href=#303 id=303 data-nosnippet>303</a>            .expect(<span class="string">"Failed to create expected RecordBatch"</span>);
-<a href=#304 id=304 data-nosnippet>304</a>            <span class="macro">assert_eq!</span>(partitioned_batches[<span class="number">0</span>].<span class="number">1</span>, expected_batch);
-<a href=#305 id=305 data-nosnippet>305</a>        }
-<a href=#306 id=306 data-nosnippet>306</a>        {
-<a href=#307 id=307 data-nosnippet>307</a>            <span class="comment">// check the second partition
-<a href=#308 id=308 data-nosnippet>308</a>            </span><span class="kw">let </span>expected_id_array = Int32Array::from(<span class="macro">vec!</span>[<span class="number">2</span>, <span class="number">2</span>]);
-<a href=#309 id=309 data-nosnippet>309</a>            <span class="kw">let </span>expected_data_array = StringArray::from(<span class="macro">vec!</span>[<span class="string">"b"</span>, <span class="string">"e"</span>]);
-<a href=#310 id=310 data-nosnippet>310</a>            <span class="kw">let </span>expected_batch = RecordBatch::try_new(arrow_schema.clone(), <span class="macro">vec!</span>[
-<a href=#311 id=311 data-nosnippet>311</a>                Arc::new(expected_id_array),
-<a href=#312 id=312 data-nosnippet>312</a>                Arc::new(expected_data_array),
-<a href=#313 id=313 data-nosnippet>313</a>            ])
-<a href=#314 id=314 data-nosnippet>314</a>            .expect(<span class="string">"Failed to create expected RecordBatch"</span>);
-<a href=#315 id=315 data-nosnippet>315</a>            <span class="macro">assert_eq!</span>(partitioned_batches[<span class="number">1</span>].<span class="number">1</span>, expected_batch);
-<a href=#316 id=316 data-nosnippet>316</a>        }
-<a href=#317 id=317 data-nosnippet>317</a>        {
-<a href=#318 id=318 data-nosnippet>318</a>            <span class="comment">// check the third partition
-<a href=#319 id=319 data-nosnippet>319</a>            </span><span class="kw">let </span>expected_id_array = Int32Array::from(<span class="macro">vec!</span>[<span class="number">3</span>, <span class="number">3</span>]);
-<a href=#320 id=320 data-nosnippet>320</a>            <span class="kw">let </span>expected_data_array = StringArray::from(<span class="macro">vec!</span>[<span class="string">"d"</span>, <span class="string">"f"</span>]);
-<a href=#321 id=321 data-nosnippet>321</a>            <span class="kw">let </span>expected_batch = RecordBatch::try_new(arrow_schema.clone(), <span class="macro">vec!</span>[
-<a href=#322 id=322 data-nosnippet>322</a>                Arc::new(expected_id_array),
-<a href=#323 id=323 data-nosnippet>323</a>                Arc::new(expected_data_array),
-<a href=#324 id=324 data-nosnippet>324</a>            ])
-<a href=#325 id=325 data-nosnippet>325</a>            .expect(<span class="string">"Failed to create expected RecordBatch"</span>);
-<a href=#326 id=326 data-nosnippet>326</a>            <span class="macro">assert_eq!</span>(partitioned_batches[<span class="number">2</span>].<span class="number">1</span>, expected_batch);
-<a href=#327 id=327 data-nosnippet>327</a>        }
-<a href=#328 id=328 data-nosnippet>328</a>
-<a href=#329 id=329 data-nosnippet>329</a>        <span class="kw">let </span>partition_values = partitioned_batches
-<a href=#330 id=330 data-nosnippet>330</a>            .iter()
-<a href=#331 id=331 data-nosnippet>331</a>            .map(|(partition_key, <span class="kw">_</span>)| partition_key.data().clone())
-<a href=#332 id=332 data-nosnippet>332</a>            .collect::&lt;Vec&lt;<span class="kw">_</span>&gt;&gt;();
-<a href=#333 id=333 data-nosnippet>333</a>        <span class="comment">// check partition value is struct(1), struct(2), struct(3)
-<a href=#334 id=334 data-nosnippet>334</a>        </span><span class="macro">assert_eq!</span>(partition_values, <span class="macro">vec!</span>[
-<a href=#335 id=335 data-nosnippet>335</a>            Struct::from_iter(<span class="macro">vec!</span>[<span class="prelude-val">Some</span>(Literal::int(<span class="number">1</span>))]),
-<a href=#336 id=336 data-nosnippet>336</a>            Struct::from_iter(<span class="macro">vec!</span>[<span class="prelude-val">Some</span>(Literal::int(<span class="number">2</span>))]),
-<a href=#337 id=337 data-nosnippet>337</a>            Struct::from_iter(<span class="macro">vec!</span>[<span class="prelude-val">Some</span>(Literal::int(<span class="number">3</span>))]),
-<a href=#338 id=338 data-nosnippet>338</a>        ]);
-<a href=#339 id=339 data-nosnippet>339</a>    }
-<a href=#340 id=340 data-nosnippet>340</a>
-<a href=#341 id=341 data-nosnippet>341</a>    <span class="attr">#[test]
-<a href=#342 id=342 data-nosnippet>342</a>    </span><span class="kw">fn </span>test_record_batch_partition_split_with_partition_column() {
-<a href=#343 id=343 data-nosnippet>343</a>        <span class="kw">use </span>arrow_array::StructArray;
-<a href=#344 id=344 data-nosnippet>344</a>        <span class="kw">use </span>arrow_schema::{Field, Schema <span class="kw">as </span>ArrowSchema};
-<a href=#345 id=345 data-nosnippet>345</a>
-<a href=#346 id=346 data-nosnippet>346</a>        <span class="kw">let </span>schema = Arc::new(
-<a href=#347 id=347 data-nosnippet>347</a>            Schema::builder()
-<a href=#348 id=348 data-nosnippet>348</a>                .with_fields(<span class="macro">vec!</span>[
-<a href=#349 id=349 data-nosnippet>349</a>                    NestedField::required(
-<a href=#350 id=350 data-nosnippet>350</a>                        <span class="number">1</span>,
-<a href=#351 id=351 data-nosnippet>351</a>                        <span class="string">"id"</span>,
-<a href=#352 id=352 data-nosnippet>352</a>                        Type::Primitive(<span class="kw">crate</span>::spec::PrimitiveType::Int),
-<a href=#353 id=353 data-nosnippet>353</a>                    )
-<a href=#354 id=354 data-nosnippet>354</a>                    .into(),
-<a href=#355 id=355 data-nosnippet>355</a>                    NestedField::required(
-<a href=#356 id=356 data-nosnippet>356</a>                        <span class="number">2</span>,
-<a href=#357 id=357 data-nosnippet>357</a>                        <span class="string">"name"</span>,
-<a href=#358 id=358 data-nosnippet>358</a>                        Type::Primitive(<span class="kw">crate</span>::spec::PrimitiveType::String),
-<a href=#359 id=359 data-nosnippet>359</a>                    )
-<a href=#360 id=360 data-nosnippet>360</a>                    .into(),
-<a href=#361 id=361 data-nosnippet>361</a>                ])
-<a href=#362 id=362 data-nosnippet>362</a>                .build()
-<a href=#363 id=363 data-nosnippet>363</a>                .unwrap(),
-<a href=#364 id=364 data-nosnippet>364</a>        );
-<a href=#365 id=365 data-nosnippet>365</a>        <span class="kw">let </span>partition_spec = Arc::new(
-<a href=#366 id=366 data-nosnippet>366</a>            PartitionSpecBuilder::new(schema.clone())
-<a href=#367 id=367 data-nosnippet>367</a>                .with_spec_id(<span class="number">1</span>)
-<a href=#368 id=368 data-nosnippet>368</a>                .add_unbound_field(UnboundPartitionField {
-<a href=#369 id=369 data-nosnippet>369</a>                    source_id: <span class="number">1</span>,
-<a href=#370 id=370 data-nosnippet>370</a>                    field_id: <span class="prelude-val">None</span>,
-<a href=#371 id=371 data-nosnippet>371</a>                    name: <span class="string">"id_bucket"</span>.to_string(),
-<a href=#372 id=372 data-nosnippet>372</a>                    transform: Transform::Identity,
-<a href=#373 id=373 data-nosnippet>373</a>                })
-<a href=#374 id=374 data-nosnippet>374</a>                .unwrap()
-<a href=#375 id=375 data-nosnippet>375</a>                .build()
-<a href=#376 id=376 data-nosnippet>376</a>                .unwrap(),
-<a href=#377 id=377 data-nosnippet>377</a>        );
-<a href=#378 id=378 data-nosnippet>378</a>
-<a href=#379 id=379 data-nosnippet>379</a>        <span class="comment">// Create input schema with _partition column
-<a href=#380 id=380 data-nosnippet>380</a>        // Note: partition field IDs start from 1000 by default
-<a href=#381 id=381 data-nosnippet>381</a>        </span><span class="kw">let </span>partition_field = Field::new(<span class="string">"id_bucket"</span>, DataType::Int32, <span class="bool-val">false</span>).with_metadata(
-<a href=#382 id=382 data-nosnippet>382</a>            HashMap::from([(PARQUET_FIELD_ID_META_KEY.to_string(), <span class="string">"1000"</span>.to_string())]),
-<a href=#383 id=383 data-nosnippet>383</a>        );
-<a href=#384 id=384 data-nosnippet>384</a>        <span class="kw">let </span>partition_struct_field = Field::new(
-<a href=#385 id=385 data-nosnippet>385</a>            PROJECTED_PARTITION_VALUE_COLUMN,
-<a href=#386 id=386 data-nosnippet>386</a>            DataType::Struct(<span class="macro">vec!</span>[partition_field.clone()].into()),
-<a href=#387 id=387 data-nosnippet>387</a>            <span class="bool-val">false</span>,
-<a href=#388 id=388 data-nosnippet>388</a>        );
-<a href=#389 id=389 data-nosnippet>389</a>
-<a href=#390 id=390 data-nosnippet>390</a>        <span class="kw">let </span>input_schema = Arc::new(ArrowSchema::new(<span class="macro">vec!</span>[
-<a href=#391 id=391 data-nosnippet>391</a>            Field::new(<span class="string">"id"</span>, DataType::Int32, <span class="bool-val">false</span>),
-<a href=#392 id=392 data-nosnippet>392</a>            Field::new(<span class="string">"name"</span>, DataType::Utf8, <span class="bool-val">false</span>),
-<a href=#393 id=393 data-nosnippet>393</a>            partition_struct_field,
-<a href=#394 id=394 data-nosnippet>394</a>        ]));
-<a href=#395 id=395 data-nosnippet>395</a>
-<a href=#396 id=396 data-nosnippet>396</a>        <span class="comment">// Create splitter expecting pre-computed partition column
-<a href=#397 id=397 data-nosnippet>397</a>        </span><span class="kw">let </span>partition_splitter = RecordBatchPartitionSplitter::try_new_with_precomputed_values(
-<a href=#398 id=398 data-nosnippet>398</a>            schema.clone(),
-<a href=#399 id=399 data-nosnippet>399</a>            partition_spec,
-<a href=#400 id=400 data-nosnippet>400</a>        )
-<a href=#401 id=401 data-nosnippet>401</a>        .expect(<span class="string">"Failed to create splitter"</span>);
-<a href=#402 id=402 data-nosnippet>402</a>
-<a href=#403 id=403 data-nosnippet>403</a>        <span class="comment">// Create test data with pre-computed partition column
-<a href=#404 id=404 data-nosnippet>404</a>        </span><span class="kw">let </span>id_array = Int32Array::from(<span class="macro">vec!</span>[<span class="number">1</span>, <span class="number">2</span>, <span class="number">1</span>, <span class="number">3</span>, <span class="number">2</span>, <span class="number">3</span>, <span class="number">1</span>]);
-<a href=#405 id=405 data-nosnippet>405</a>        <span class="kw">let </span>data_array = StringArray::from(<span class="macro">vec!</span>[<span class="string">"a"</span>, <span class="string">"b"</span>, <span class="string">"c"</span>, <span class="string">"d"</span>, <span class="string">"e"</span>, <span class="string">"f"</span>, <span class="string">"g"</span>]);
-<a href=#406 id=406 data-nosnippet>406</a>
-<a href=#407 id=407 data-nosnippet>407</a>        <span class="comment">// Create partition column (same values as id for Identity transform)
-<a href=#408 id=408 data-nosnippet>408</a>        </span><span class="kw">let </span>partition_values = Int32Array::from(<span class="macro">vec!</span>[<span class="number">1</span>, <span class="number">2</span>, <span class="number">1</span>, <span class="number">3</span>, <span class="number">2</span>, <span class="number">3</span>, <span class="number">1</span>]);
-<a href=#409 id=409 data-nosnippet>409</a>        <span class="kw">let </span>partition_struct = StructArray::from(<span class="macro">vec!</span>[(
-<a href=#410 id=410 data-nosnippet>410</a>            Arc::new(partition_field),
-<a href=#411 id=411 data-nosnippet>411</a>            Arc::new(partition_values) <span class="kw">as </span>ArrayRef,
-<a href=#412 id=412 data-nosnippet>412</a>        )]);
-<a href=#413 id=413 data-nosnippet>413</a>
-<a href=#414 id=414 data-nosnippet>414</a>        <span class="kw">let </span>batch = RecordBatch::try_new(input_schema.clone(), <span class="macro">vec!</span>[
-<a href=#415 id=415 data-nosnippet>415</a>            Arc::new(id_array),
-<a href=#416 id=416 data-nosnippet>416</a>            Arc::new(data_array),
-<a href=#417 id=417 data-nosnippet>417</a>            Arc::new(partition_struct),
-<a href=#418 id=418 data-nosnippet>418</a>        ])
-<a href=#419 id=419 data-nosnippet>419</a>        .expect(<span class="string">"Failed to create RecordBatch"</span>);
-<a href=#420 id=420 data-nosnippet>420</a>
-<a href=#421 id=421 data-nosnippet>421</a>        <span class="comment">// Split using the pre-computed partition column
-<a href=#422 id=422 data-nosnippet>422</a>        </span><span class="kw">let </span><span class="kw-2">mut </span>partitioned_batches = partition_splitter
-<a href=#423 id=423 data-nosnippet>423</a>            .split(<span class="kw-2">&amp;</span>batch)
-<a href=#424 id=424 data-nosnippet>424</a>            .expect(<span class="string">"Failed to split RecordBatch"</span>);
-<a href=#425 id=425 data-nosnippet>425</a>
-<a href=#426 id=426 data-nosnippet>426</a>        partitioned_batches.sort_by_key(|(partition_key, <span class="kw">_</span>)| {
-<a href=#427 id=427 data-nosnippet>427</a>            <span class="kw">if let </span>PrimitiveLiteral::Int(i) = partition_key.data().fields()[<span class="number">0</span>]
-<a href=#428 id=428 data-nosnippet>428</a>                .as_ref()
-<a href=#429 id=429 data-nosnippet>429</a>                .unwrap()
-<a href=#430 id=430 data-nosnippet>430</a>                .as_primitive_literal()
+<a href=#288 id=288 data-nosnippet>288</a>                .as_primitive_literal()
+<a href=#289 id=289 data-nosnippet>289</a>                .unwrap()
+<a href=#290 id=290 data-nosnippet>290</a>            {
+<a href=#291 id=291 data-nosnippet>291</a>                i
+<a href=#292 id=292 data-nosnippet>292</a>            } <span class="kw">else </span>{
+<a href=#293 id=293 data-nosnippet>293</a>                <span class="macro">panic!</span>(<span class="string">"The partition value is not a int"</span>);
+<a href=#294 id=294 data-nosnippet>294</a>            }
+<a href=#295 id=295 data-nosnippet>295</a>        });
+<a href=#296 id=296 data-nosnippet>296</a>        <span class="macro">assert_eq!</span>(partitioned_batches.len(), <span class="number">3</span>);
+<a href=#297 id=297 data-nosnippet>297</a>        {
+<a href=#298 id=298 data-nosnippet>298</a>            <span class="comment">// check the first partition
+<a href=#299 id=299 data-nosnippet>299</a>            </span><span class="kw">let </span>expected_id_array = Int32Array::from(<span class="macro">vec!</span>[<span class="number">1</span>, <span class="number">1</span>, <span class="number">1</span>]);
+<a href=#300 id=300 data-nosnippet>300</a>            <span class="kw">let </span>expected_data_array = StringArray::from(<span class="macro">vec!</span>[<span class="string">"a"</span>, <span class="string">"c"</span>, <span class="string">"g"</span>]);
+<a href=#301 id=301 data-nosnippet>301</a>            <span class="kw">let </span>expected_batch = RecordBatch::try_new(arrow_schema.clone(), <span class="macro">vec!</span>[
+<a href=#302 id=302 data-nosnippet>302</a>                Arc::new(expected_id_array),
+<a href=#303 id=303 data-nosnippet>303</a>                Arc::new(expected_data_array),
+<a href=#304 id=304 data-nosnippet>304</a>            ])
+<a href=#305 id=305 data-nosnippet>305</a>            .expect(<span class="string">"Failed to create expected RecordBatch"</span>);
+<a href=#306 id=306 data-nosnippet>306</a>            <span class="macro">assert_eq!</span>(partitioned_batches[<span class="number">0</span>].<span class="number">1</span>, expected_batch);
+<a href=#307 id=307 data-nosnippet>307</a>        }
+<a href=#308 id=308 data-nosnippet>308</a>        {
+<a href=#309 id=309 data-nosnippet>309</a>            <span class="comment">// check the second partition
+<a href=#310 id=310 data-nosnippet>310</a>            </span><span class="kw">let </span>expected_id_array = Int32Array::from(<span class="macro">vec!</span>[<span class="number">2</span>, <span class="number">2</span>]);
+<a href=#311 id=311 data-nosnippet>311</a>            <span class="kw">let </span>expected_data_array = StringArray::from(<span class="macro">vec!</span>[<span class="string">"b"</span>, <span class="string">"e"</span>]);
+<a href=#312 id=312 data-nosnippet>312</a>            <span class="kw">let </span>expected_batch = RecordBatch::try_new(arrow_schema.clone(), <span class="macro">vec!</span>[
+<a href=#313 id=313 data-nosnippet>313</a>                Arc::new(expected_id_array),
+<a href=#314 id=314 data-nosnippet>314</a>                Arc::new(expected_data_array),
+<a href=#315 id=315 data-nosnippet>315</a>            ])
+<a href=#316 id=316 data-nosnippet>316</a>            .expect(<span class="string">"Failed to create expected RecordBatch"</span>);
+<a href=#317 id=317 data-nosnippet>317</a>            <span class="macro">assert_eq!</span>(partitioned_batches[<span class="number">1</span>].<span class="number">1</span>, expected_batch);
+<a href=#318 id=318 data-nosnippet>318</a>        }
+<a href=#319 id=319 data-nosnippet>319</a>        {
+<a href=#320 id=320 data-nosnippet>320</a>            <span class="comment">// check the third partition
+<a href=#321 id=321 data-nosnippet>321</a>            </span><span class="kw">let </span>expected_id_array = Int32Array::from(<span class="macro">vec!</span>[<span class="number">3</span>, <span class="number">3</span>]);
+<a href=#322 id=322 data-nosnippet>322</a>            <span class="kw">let </span>expected_data_array = StringArray::from(<span class="macro">vec!</span>[<span class="string">"d"</span>, <span class="string">"f"</span>]);
+<a href=#323 id=323 data-nosnippet>323</a>            <span class="kw">let </span>expected_batch = RecordBatch::try_new(arrow_schema.clone(), <span class="macro">vec!</span>[
+<a href=#324 id=324 data-nosnippet>324</a>                Arc::new(expected_id_array),
+<a href=#325 id=325 data-nosnippet>325</a>                Arc::new(expected_data_array),
+<a href=#326 id=326 data-nosnippet>326</a>            ])
+<a href=#327 id=327 data-nosnippet>327</a>            .expect(<span class="string">"Failed to create expected RecordBatch"</span>);
+<a href=#328 id=328 data-nosnippet>328</a>            <span class="macro">assert_eq!</span>(partitioned_batches[<span class="number">2</span>].<span class="number">1</span>, expected_batch);
+<a href=#329 id=329 data-nosnippet>329</a>        }
+<a href=#330 id=330 data-nosnippet>330</a>
+<a href=#331 id=331 data-nosnippet>331</a>        <span class="kw">let </span>partition_values = partitioned_batches
+<a href=#332 id=332 data-nosnippet>332</a>            .iter()
+<a href=#333 id=333 data-nosnippet>333</a>            .map(|(partition_key, <span class="kw">_</span>)| partition_key.data().clone())
+<a href=#334 id=334 data-nosnippet>334</a>            .collect::&lt;Vec&lt;<span class="kw">_</span>&gt;&gt;();
+<a href=#335 id=335 data-nosnippet>335</a>        <span class="comment">// check partition value is struct(1), struct(2), struct(3)
+<a href=#336 id=336 data-nosnippet>336</a>        </span><span class="macro">assert_eq!</span>(partition_values, <span class="macro">vec!</span>[
+<a href=#337 id=337 data-nosnippet>337</a>            Struct::from_iter(<span class="macro">vec!</span>[<span class="prelude-val">Some</span>(Literal::int(<span class="number">1</span>))]),
+<a href=#338 id=338 data-nosnippet>338</a>            Struct::from_iter(<span class="macro">vec!</span>[<span class="prelude-val">Some</span>(Literal::int(<span class="number">2</span>))]),
+<a href=#339 id=339 data-nosnippet>339</a>            Struct::from_iter(<span class="macro">vec!</span>[<span class="prelude-val">Some</span>(Literal::int(<span class="number">3</span>))]),
+<a href=#340 id=340 data-nosnippet>340</a>        ]);
+<a href=#341 id=341 data-nosnippet>341</a>    }
+<a href=#342 id=342 data-nosnippet>342</a>
+<a href=#343 id=343 data-nosnippet>343</a>    <span class="attr">#[test]
+<a href=#344 id=344 data-nosnippet>344</a>    </span><span class="kw">fn </span>test_record_batch_partition_split_with_partition_column() {
+<a href=#345 id=345 data-nosnippet>345</a>        <span class="kw">use </span>arrow_array::StructArray;
+<a href=#346 id=346 data-nosnippet>346</a>        <span class="kw">use </span>arrow_schema::{Field, Schema <span class="kw">as </span>ArrowSchema};
+<a href=#347 id=347 data-nosnippet>347</a>
+<a href=#348 id=348 data-nosnippet>348</a>        <span class="kw">let </span>schema = Arc::new(
+<a href=#349 id=349 data-nosnippet>349</a>            Schema::builder()
+<a href=#350 id=350 data-nosnippet>350</a>                .with_fields(<span class="macro">vec!</span>[
+<a href=#351 id=351 data-nosnippet>351</a>                    NestedField::required(
+<a href=#352 id=352 data-nosnippet>352</a>                        <span class="number">1</span>,
+<a href=#353 id=353 data-nosnippet>353</a>                        <span class="string">"id"</span>,
+<a href=#354 id=354 data-nosnippet>354</a>                        Type::Primitive(<span class="kw">crate</span>::spec::PrimitiveType::Int),
+<a href=#355 id=355 data-nosnippet>355</a>                    )
+<a href=#356 id=356 data-nosnippet>356</a>                    .into(),
+<a href=#357 id=357 data-nosnippet>357</a>                    NestedField::required(
+<a href=#358 id=358 data-nosnippet>358</a>                        <span class="number">2</span>,
+<a href=#359 id=359 data-nosnippet>359</a>                        <span class="string">"name"</span>,
+<a href=#360 id=360 data-nosnippet>360</a>                        Type::Primitive(<span class="kw">crate</span>::spec::PrimitiveType::String),
+<a href=#361 id=361 data-nosnippet>361</a>                    )
+<a href=#362 id=362 data-nosnippet>362</a>                    .into(),
+<a href=#363 id=363 data-nosnippet>363</a>                ])
+<a href=#364 id=364 data-nosnippet>364</a>                .build()
+<a href=#365 id=365 data-nosnippet>365</a>                .unwrap(),
+<a href=#366 id=366 data-nosnippet>366</a>        );
+<a href=#367 id=367 data-nosnippet>367</a>        <span class="kw">let </span>partition_spec = Arc::new(
+<a href=#368 id=368 data-nosnippet>368</a>            PartitionSpecBuilder::new(schema.clone())
+<a href=#369 id=369 data-nosnippet>369</a>                .with_spec_id(<span class="number">1</span>)
+<a href=#370 id=370 data-nosnippet>370</a>                .add_unbound_field(UnboundPartitionField {
+<a href=#371 id=371 data-nosnippet>371</a>                    source_id: <span class="number">1</span>,
+<a href=#372 id=372 data-nosnippet>372</a>                    field_id: <span class="prelude-val">None</span>,
+<a href=#373 id=373 data-nosnippet>373</a>                    name: <span class="string">"id_bucket"</span>.to_string(),
+<a href=#374 id=374 data-nosnippet>374</a>                    transform: Transform::Identity,
+<a href=#375 id=375 data-nosnippet>375</a>                })
+<a href=#376 id=376 data-nosnippet>376</a>                .unwrap()
+<a href=#377 id=377 data-nosnippet>377</a>                .build()
+<a href=#378 id=378 data-nosnippet>378</a>                .unwrap(),
+<a href=#379 id=379 data-nosnippet>379</a>        );
+<a href=#380 id=380 data-nosnippet>380</a>
+<a href=#381 id=381 data-nosnippet>381</a>        <span class="comment">// Create input schema with _partition column
+<a href=#382 id=382 data-nosnippet>382</a>        // Note: partition field IDs start from 1000 by default
+<a href=#383 id=383 data-nosnippet>383</a>        </span><span class="kw">let </span>partition_field = Field::new(<span class="string">"id_bucket"</span>, DataType::Int32, <span class="bool-val">false</span>).with_metadata(
+<a href=#384 id=384 data-nosnippet>384</a>            HashMap::from([(PARQUET_FIELD_ID_META_KEY.to_string(), <span class="string">"1000"</span>.to_string())]),
+<a href=#385 id=385 data-nosnippet>385</a>        );
+<a href=#386 id=386 data-nosnippet>386</a>        <span class="kw">let </span>partition_struct_field = Field::new(
+<a href=#387 id=387 data-nosnippet>387</a>            PROJECTED_PARTITION_VALUE_COLUMN,
+<a href=#388 id=388 data-nosnippet>388</a>            DataType::Struct(<span class="macro">vec!</span>[partition_field.clone()].into()),
+<a href=#389 id=389 data-nosnippet>389</a>            <span class="bool-val">false</span>,
+<a href=#390 id=390 data-nosnippet>390</a>        );
+<a href=#391 id=391 data-nosnippet>391</a>
+<a href=#392 id=392 data-nosnippet>392</a>        <span class="kw">let </span>input_schema = Arc::new(ArrowSchema::new(<span class="macro">vec!</span>[
+<a href=#393 id=393 data-nosnippet>393</a>            Field::new(<span class="string">"id"</span>, DataType::Int32, <span class="bool-val">false</span>),
+<a href=#394 id=394 data-nosnippet>394</a>            Field::new(<span class="string">"name"</span>, DataType::Utf8, <span class="bool-val">false</span>),
+<a href=#395 id=395 data-nosnippet>395</a>            partition_struct_field,
+<a href=#396 id=396 data-nosnippet>396</a>        ]));
+<a href=#397 id=397 data-nosnippet>397</a>
+<a href=#398 id=398 data-nosnippet>398</a>        <span class="comment">// Create splitter expecting pre-computed partition column
+<a href=#399 id=399 data-nosnippet>399</a>        </span><span class="kw">let </span>partition_splitter = RecordBatchPartitionSplitter::try_new_with_precomputed_values(
+<a href=#400 id=400 data-nosnippet>400</a>            schema.clone(),
+<a href=#401 id=401 data-nosnippet>401</a>            partition_spec,
+<a href=#402 id=402 data-nosnippet>402</a>        )
+<a href=#403 id=403 data-nosnippet>403</a>        .expect(<span class="string">"Failed to create splitter"</span>);
+<a href=#404 id=404 data-nosnippet>404</a>
+<a href=#405 id=405 data-nosnippet>405</a>        <span class="comment">// Create test data with pre-computed partition column
+<a href=#406 id=406 data-nosnippet>406</a>        </span><span class="kw">let </span>id_array = Int32Array::from(<span class="macro">vec!</span>[<span class="number">1</span>, <span class="number">2</span>, <span class="number">1</span>, <span class="number">3</span>, <span class="number">2</span>, <span class="number">3</span>, <span class="number">1</span>]);
+<a href=#407 id=407 data-nosnippet>407</a>        <span class="kw">let </span>data_array = StringArray::from(<span class="macro">vec!</span>[<span class="string">"a"</span>, <span class="string">"b"</span>, <span class="string">"c"</span>, <span class="string">"d"</span>, <span class="string">"e"</span>, <span class="string">"f"</span>, <span class="string">"g"</span>]);
+<a href=#408 id=408 data-nosnippet>408</a>
+<a href=#409 id=409 data-nosnippet>409</a>        <span class="comment">// Create partition column (same values as id for Identity transform)
+<a href=#410 id=410 data-nosnippet>410</a>        </span><span class="kw">let </span>partition_values = Int32Array::from(<span class="macro">vec!</span>[<span class="number">1</span>, <span class="number">2</span>, <span class="number">1</span>, <span class="number">3</span>, <span class="number">2</span>, <span class="number">3</span>, <span class="number">1</span>]);
+<a href=#411 id=411 data-nosnippet>411</a>        <span class="kw">let </span>partition_struct = StructArray::from(<span class="macro">vec!</span>[(
+<a href=#412 id=412 data-nosnippet>412</a>            Arc::new(partition_field),
+<a href=#413 id=413 data-nosnippet>413</a>            Arc::new(partition_values) <span class="kw">as </span>ArrayRef,
+<a href=#414 id=414 data-nosnippet>414</a>        )]);
+<a href=#415 id=415 data-nosnippet>415</a>
+<a href=#416 id=416 data-nosnippet>416</a>        <span class="kw">let </span>batch = RecordBatch::try_new(input_schema.clone(), <span class="macro">vec!</span>[
+<a href=#417 id=417 data-nosnippet>417</a>            Arc::new(id_array),
+<a href=#418 id=418 data-nosnippet>418</a>            Arc::new(data_array),
+<a href=#419 id=419 data-nosnippet>419</a>            Arc::new(partition_struct),
+<a href=#420 id=420 data-nosnippet>420</a>        ])
+<a href=#421 id=421 data-nosnippet>421</a>        .expect(<span class="string">"Failed to create RecordBatch"</span>);
+<a href=#422 id=422 data-nosnippet>422</a>
+<a href=#423 id=423 data-nosnippet>423</a>        <span class="comment">// Split using the pre-computed partition column
+<a href=#424 id=424 data-nosnippet>424</a>        </span><span class="kw">let </span><span class="kw-2">mut </span>partitioned_batches = partition_splitter
+<a href=#425 id=425 data-nosnippet>425</a>            .split(<span class="kw-2">&amp;</span>batch)
+<a href=#426 id=426 data-nosnippet>426</a>            .expect(<span class="string">"Failed to split RecordBatch"</span>);
+<a href=#427 id=427 data-nosnippet>427</a>
+<a href=#428 id=428 data-nosnippet>428</a>        partitioned_batches.sort_by_key(|(partition_key, <span class="kw">_</span>)| {
+<a href=#429 id=429 data-nosnippet>429</a>            <span class="kw">if let </span>PrimitiveLiteral::Int(i) = partition_key.data().fields()[<span class="number">0</span>]
+<a href=#430 id=430 data-nosnippet>430</a>                .as_ref()
 <a href=#431 id=431 data-nosnippet>431</a>                .unwrap()
-<a href=#432 id=432 data-nosnippet>432</a>            {
-<a href=#433 id=433 data-nosnippet>433</a>                i
-<a href=#434 id=434 data-nosnippet>434</a>            } <span class="kw">else </span>{
-<a href=#435 id=435 data-nosnippet>435</a>                <span class="macro">panic!</span>(<span class="string">"The partition value is not a int"</span>);
-<a href=#436 id=436 data-nosnippet>436</a>            }
-<a href=#437 id=437 data-nosnippet>437</a>        });
-<a href=#438 id=438 data-nosnippet>438</a>
-<a href=#439 id=439 data-nosnippet>439</a>        <span class="macro">assert_eq!</span>(partitioned_batches.len(), <span class="number">3</span>);
+<a href=#432 id=432 data-nosnippet>432</a>                .as_primitive_literal()
+<a href=#433 id=433 data-nosnippet>433</a>                .unwrap()
+<a href=#434 id=434 data-nosnippet>434</a>            {
+<a href=#435 id=435 data-nosnippet>435</a>                i
+<a href=#436 id=436 data-nosnippet>436</a>            } <span class="kw">else </span>{
+<a href=#437 id=437 data-nosnippet>437</a>                <span class="macro">panic!</span>(<span class="string">"The partition value is not a int"</span>);
+<a href=#438 id=438 data-nosnippet>438</a>            }
+<a href=#439 id=439 data-nosnippet>439</a>        });
 <a href=#440 id=440 data-nosnippet>440</a>
-<a href=#441 id=441 data-nosnippet>441</a>        <span class="comment">// Helper to extract id and name values from a batch
-<a href=#442 id=442 data-nosnippet>442</a>        </span><span class="kw">let </span>extract_values = |batch: <span class="kw-2">&amp;</span>RecordBatch| -&gt; (Vec&lt;i32&gt;, Vec&lt;String&gt;) {
-<a href=#443 id=443 data-nosnippet>443</a>            <span class="kw">let </span>id_col = batch
-<a href=#444 id=444 data-nosnippet>444</a>                .column(<span class="number">0</span>)
-<a href=#445 id=445 data-nosnippet>445</a>                .as_any()
-<a href=#446 id=446 data-nosnippet>446</a>                .downcast_ref::&lt;Int32Array&gt;()
-<a href=#447 id=447 data-nosnippet>447</a>                .unwrap();
-<a href=#448 id=448 data-nosnippet>448</a>            <span class="kw">let </span>name_col = batch
-<a href=#449 id=449 data-nosnippet>449</a>                .column(<span class="number">1</span>)
-<a href=#450 id=450 data-nosnippet>450</a>                .as_any()
-<a href=#451 id=451 data-nosnippet>451</a>                .downcast_ref::&lt;StringArray&gt;()
-<a href=#452 id=452 data-nosnippet>452</a>                .unwrap();
-<a href=#453 id=453 data-nosnippet>453</a>            (
-<a href=#454 id=454 data-nosnippet>454</a>                id_col.values().to_vec(),
-<a href=#455 id=455 data-nosnippet>455</a>                name_col.iter().map(|s| s.unwrap().to_string()).collect(),
-<a href=#456 id=456 data-nosnippet>456</a>            )
-<a href=#457 id=457 data-nosnippet>457</a>        };
-<a href=#458 id=458 data-nosnippet>458</a>
-<a href=#459 id=459 data-nosnippet>459</a>        <span class="comment">// Verify partition 1: id=1, names=["a", "c", "g"]
-<a href=#460 id=460 data-nosnippet>460</a>        </span><span class="kw">let </span>(key, batch) = <span class="kw-2">&amp;</span>partitioned_batches[<span class="number">0</span>];
-<a href=#461 id=461 data-nosnippet>461</a>        <span class="macro">assert_eq!</span>(key.data(), <span class="kw-2">&amp;</span>Struct::from_iter(<span class="macro">vec!</span>[<span class="prelude-val">Some</span>(Literal::int(<span class="number">1</span>))]));
-<a href=#462 id=462 data-nosnippet>462</a>        <span class="kw">let </span>(ids, names) = extract_values(batch);
-<a href=#463 id=463 data-nosnippet>463</a>        <span class="macro">assert_eq!</span>(ids, <span class="macro">vec!</span>[<span class="number">1</span>, <span class="number">1</span>, <span class="number">1</span>]);
-<a href=#464 id=464 data-nosnippet>464</a>        <span class="macro">assert_eq!</span>(names, <span class="macro">vec!</span>[<span class="string">"a"</span>, <span class="string">"c"</span>, <span class="string">"g"</span>]);
-<a href=#465 id=465 data-nosnippet>465</a>
-<a href=#466 id=466 data-nosnippet>466</a>        <span class="comment">// Verify partition 2: id=2, names=["b", "e"]
-<a href=#467 id=467 data-nosnippet>467</a>        </span><span class="kw">let </span>(key, batch) = <span class="kw-2">&amp;</span>partitioned_batches[<span class="number">1</span>];
-<a href=#468 id=468 data-nosnippet>468</a>        <span class="macro">assert_eq!</span>(key.data(), <span class="kw-2">&amp;</span>Struct::from_iter(<span class="macro">vec!</span>[<span class="prelude-val">Some</span>(Literal::int(<span class="number">2</span>))]));
-<a href=#469 id=469 data-nosnippet>469</a>        <span class="kw">let </span>(ids, names) = extract_values(batch);
-<a href=#470 id=470 data-nosnippet>470</a>        <span class="macro">assert_eq!</span>(ids, <span class="macro">vec!</span>[<span class="number">2</span>, <span class="number">2</span>]);
-<a href=#471 id=471 data-nosnippet>471</a>        <span class="macro">assert_eq!</span>(names, <span class="macro">vec!</span>[<span class="string">"b"</span>, <span class="string">"e"</span>]);
-<a href=#472 id=472 data-nosnippet>472</a>
-<a href=#473 id=473 data-nosnippet>473</a>        <span class="comment">// Verify partition 3: id=3, names=["d", "f"]
-<a href=#474 id=474 data-nosnippet>474</a>        </span><span class="kw">let </span>(key, batch) = <span class="kw-2">&amp;</span>partitioned_batches[<span class="number">2</span>];
-<a href=#475 id=475 data-nosnippet>475</a>        <span class="macro">assert_eq!</span>(key.data(), <span class="kw-2">&amp;</span>Struct::from_iter(<span class="macro">vec!</span>[<span class="prelude-val">Some</span>(Literal::int(<span class="number">3</span>))]));
-<a href=#476 id=476 data-nosnippet>476</a>        <span class="kw">let </span>(ids, names) = extract_values(batch);
-<a href=#477 id=477 data-nosnippet>477</a>        <span class="macro">assert_eq!</span>(ids, <span class="macro">vec!</span>[<span class="number">3</span>, <span class="number">3</span>]);
-<a href=#478 id=478 data-nosnippet>478</a>        <span class="macro">assert_eq!</span>(names, <span class="macro">vec!</span>[<span class="string">"d"</span>, <span class="string">"f"</span>]);
-<a href=#479 id=479 data-nosnippet>479</a>    }
-<a href=#480 id=480 data-nosnippet>480</a>}
+<a href=#441 id=441 data-nosnippet>441</a>        <span class="macro">assert_eq!</span>(partitioned_batches.len(), <span class="number">3</span>);
+<a href=#442 id=442 data-nosnippet>442</a>
+<a href=#443 id=443 data-nosnippet>443</a>        <span class="comment">// Helper to extract id and name values from a batch
+<a href=#444 id=444 data-nosnippet>444</a>        </span><span class="kw">let </span>extract_values = |batch: <span class="kw-2">&amp;</span>RecordBatch| -&gt; (Vec&lt;i32&gt;, Vec&lt;String&gt;) {
+<a href=#445 id=445 data-nosnippet>445</a>            <span class="kw">let </span>id_col = batch
+<a href=#446 id=446 data-nosnippet>446</a>                .column(<span class="number">0</span>)
+<a href=#447 id=447 data-nosnippet>447</a>                .as_any()
+<a href=#448 id=448 data-nosnippet>448</a>                .downcast_ref::&lt;Int32Array&gt;()
+<a href=#449 id=449 data-nosnippet>449</a>                .unwrap();
+<a href=#450 id=450 data-nosnippet>450</a>            <span class="kw">let </span>name_col = batch
+<a href=#451 id=451 data-nosnippet>451</a>                .column(<span class="number">1</span>)
+<a href=#452 id=452 data-nosnippet>452</a>                .as_any()
+<a href=#453 id=453 data-nosnippet>453</a>                .downcast_ref::&lt;StringArray&gt;()
+<a href=#454 id=454 data-nosnippet>454</a>                .unwrap();
+<a href=#455 id=455 data-nosnippet>455</a>            (
+<a href=#456 id=456 data-nosnippet>456</a>                id_col.values().to_vec(),
+<a href=#457 id=457 data-nosnippet>457</a>                name_col.iter().map(|s| s.unwrap().to_string()).collect(),
+<a href=#458 id=458 data-nosnippet>458</a>            )
+<a href=#459 id=459 data-nosnippet>459</a>        };
+<a href=#460 id=460 data-nosnippet>460</a>
+<a href=#461 id=461 data-nosnippet>461</a>        <span class="comment">// Verify partition 1: id=1, names=["a", "c", "g"]
+<a href=#462 id=462 data-nosnippet>462</a>        </span><span class="kw">let </span>(key, batch) = <span class="kw-2">&amp;</span>partitioned_batches[<span class="number">0</span>];
+<a href=#463 id=463 data-nosnippet>463</a>        <span class="macro">assert_eq!</span>(key.data(), <span class="kw-2">&amp;</span>Struct::from_iter(<span class="macro">vec!</span>[<span class="prelude-val">Some</span>(Literal::int(<span class="number">1</span>))]));
+<a href=#464 id=464 data-nosnippet>464</a>        <span class="kw">let </span>(ids, names) = extract_values(batch);
+<a href=#465 id=465 data-nosnippet>465</a>        <span class="macro">assert_eq!</span>(ids, <span class="macro">vec!</span>[<span class="number">1</span>, <span class="number">1</span>, <span class="number">1</span>]);
+<a href=#466 id=466 data-nosnippet>466</a>        <span class="macro">assert_eq!</span>(names, <span class="macro">vec!</span>[<span class="string">"a"</span>, <span class="string">"c"</span>, <span class="string">"g"</span>]);
+<a href=#467 id=467 data-nosnippet>467</a>
+<a href=#468 id=468 data-nosnippet>468</a>        <span class="comment">// Verify partition 2: id=2, names=["b", "e"]
+<a href=#469 id=469 data-nosnippet>469</a>        </span><span class="kw">let </span>(key, batch) = <span class="kw-2">&amp;</span>partitioned_batches[<span class="number">1</span>];
+<a href=#470 id=470 data-nosnippet>470</a>        <span class="macro">assert_eq!</span>(key.data(), <span class="kw-2">&amp;</span>Struct::from_iter(<span class="macro">vec!</span>[<span class="prelude-val">Some</span>(Literal::int(<span class="number">2</span>))]));
+<a href=#471 id=471 data-nosnippet>471</a>        <span class="kw">let </span>(ids, names) = extract_values(batch);
+<a href=#472 id=472 data-nosnippet>472</a>        <span class="macro">assert_eq!</span>(ids, <span class="macro">vec!</span>[<span class="number">2</span>, <span class="number">2</span>]);
+<a href=#473 id=473 data-nosnippet>473</a>        <span class="macro">assert_eq!</span>(names, <span class="macro">vec!</span>[<span class="string">"b"</span>, <span class="string">"e"</span>]);
+<a href=#474 id=474 data-nosnippet>474</a>
+<a href=#475 id=475 data-nosnippet>475</a>        <span class="comment">// Verify partition 3: id=3, names=["d", "f"]
+<a href=#476 id=476 data-nosnippet>476</a>        </span><span class="kw">let </span>(key, batch) = <span class="kw-2">&amp;</span>partitioned_batches[<span class="number">2</span>];
+<a href=#477 id=477 data-nosnippet>477</a>        <span class="macro">assert_eq!</span>(key.data(), <span class="kw-2">&amp;</span>Struct::from_iter(<span class="macro">vec!</span>[<span class="prelude-val">Some</span>(Literal::int(<span class="number">3</span>))]));
+<a href=#478 id=478 data-nosnippet>478</a>        <span class="kw">let </span>(ids, names) = extract_values(batch);
+<a href=#479 id=479 data-nosnippet>479</a>        <span class="macro">assert_eq!</span>(ids, <span class="macro">vec!</span>[<span class="number">3</span>, <span class="number">3</span>]);
+<a href=#480 id=480 data-nosnippet>480</a>        <span class="macro">assert_eq!</span>(names, <span class="macro">vec!</span>[<span class="string">"d"</span>, <span class="string">"f"</span>]);
+<a href=#481 id=481 data-nosnippet>481</a>    }
+<a href=#482 id=482 data-nosnippet>482</a>}
 </code></pre></div></section></main></body></html>
\ No newline at end of file