[GitHub] [arrow-site] alamb commented on a diff in pull request #247: ARROW-17907: [Website] Arbitrarily Nested Data in Parquet and Arrow: Part 3: Encoding Arbitrary Structs of Lists, Lists of Structs

[GitBox <gi...@apache.org>]

alamb commented on code in PR #247:
URL: https://github.com/apache/arrow-site/pull/247#discussion_r985096328


##########
_posts/2022-10-01-arrow-parquet-encoding-part-3.md:
##########
@@ -0,0 +1,169 @@
+---
+layout: post
+title: Arrow and Parquet Part 3: Arbitrary Nesting with Lists of Structs and Structs of Lists
+date: "2022-10-01 00:00:00"
+author: tustvold, alamb
+categories: [parquet, arrow]
+---
+<!--
+{% comment %}
+Licensed to the Apache Software Foundation (ASF) under one or more
+contributor license agreements.  See the NOTICE file distributed with
+this work for additional information regarding copyright ownership.
+The ASF licenses this file to you under the Apache License, Version 2.0
+(the "License"); you may not use this file except in compliance with
+the License.  You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+{% endcomment %}
+-->
+
+## Introduction
+
+This is the third of a three part series exploring how projects such as [Rust Apache Arrow](https://github.com/apache/arrow-rs) support conversion between [Apache Arrow](https://arrow.apache.org/) for in memory processing and [Apache Parquet](https://parquet.apache.org/) for efficient storage. This post covers how to combine the `Struct` and `List` types described in the previous posts for arbitrary nesting.
+
+
+[Apache Arrow](https://arrow.apache.org/) is an open, language-independent columnar memory format for flat and hierarchical data, organized for efficient analytic operations. [Apache Parquet](https://parquet.apache.org/) is an open, column-oriented data file format designed for very efficient data encoding and retrieval.
+
+
+# Structs with Lists
+
+
+```json
+{                     <-- First record
+  “a”: [1],           <-- top-level field a containing list of integers
+  “b”: [              <-- top-level field b containing list of structures
+    {                 <-- list element of b containing two field b1 and b2
+      “b1”: 1         <-- b1 is always provided (not null)
+    },
+    {
+      “b1”: 1,
+      “b2”: [         <-- b2 contains list of integers
+        3, 4          <-- list elements of b.b2 always provided (not null)
+      ]
+    }
+  ]
+}
+{
+  “b”: [              <-- b is always provided (not null)
+    {
+      “b1”: 2
+    },
+  ]
+}
+{
+  “a”: [null, null],  <-- list elements of a are nullable
+  “b”: [null]         <-- list elements of b are nullable
+}
+```
+
+Documents of this format could be stored in this arrow schema
+
+```text
+Field(name: “a”, nullable: true, datatype: List(
+  Field(name: “element”, nullable: true, datatype: Int32),
+)
+Field(name: “b”), nullable: false, datatype: List(
+  Field(name: “element”, nullable: true, datatype: Struct[
+    Field(name: “b2”, nullable: false, datatype: Int32),
+    Field(name: “c2”, nullable: true, datatype: List(
+      Field(name: “element”, nullable: false, datatype: Int32)
+    ))
+  ])
+))
+```
+
+Documents of this format could be stored in this parquet schema
+
+```text
+message schema {
+  optional group a (LIST) {
+    repeated group list {
+      optional int32 element;
+    }
+  }
+  required group b (LIST) {
+    repeated group list {
+      optional group element {
+        required int32 c1;
+        optional group c2 (LIST) {
+          repeated group list {
+            required int32 element;
+          }
+        }
+      }
+    }
+  }
+}
+```
+
+As explained previously, Arrow chooses to represent this in a hierarchical fashion. To achieve this it stores a list of monotonically increasing integers called offsets in the parent ListArray, and stores all the values that appear in the lists in a single child array. Each consecutive pair of elements in this offset array identifies a slice of the child array for that array index.
+
+```text

Review Comment:
   I think making this into a diagram would be good -- I just haven't had a chance



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@arrow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org