[GitHub] [arrow-rs] tustvold commented on a diff in pull request #3212: Support StructArray in Row Format (#3159)

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

tustvold commented on code in PR #3212:
URL: https://github.com/apache/arrow-rs/pull/3212#discussion_r1033972417


##########
arrow/src/row/mod.rs:
##########
@@ -872,35 +980,59 @@ fn encode_column(
     out: &mut Rows,
     column: &ArrayRef,
     opts: SortOptions,
-    dictionary: Option<&[Option<&[u8]>]>,
+    encoder: &Encoder<'_>,
 ) {
-    downcast_primitive_array! {
-        column => fixed::encode(out, column, opts),
-        DataType::Null => {}
-        DataType::Boolean => fixed::encode(out, as_boolean_array(column), opts),
-        DataType::Binary => {
-            variable::encode(out, as_generic_binary_array::<i32>(column).iter(), opts)
+    match encoder {
+        Encoder::Stateless => {
+            downcast_primitive_array! {
+                column => fixed::encode(out, column, opts),
+                DataType::Null => {}
+                DataType::Boolean => fixed::encode(out, as_boolean_array(column), opts),
+                DataType::Binary => {
+                    variable::encode(out, as_generic_binary_array::<i32>(column).iter(), opts)
+                }
+                DataType::LargeBinary => {
+                    variable::encode(out, as_generic_binary_array::<i64>(column).iter(), opts)
+                }
+                DataType::Utf8 => variable::encode(
+                    out,
+                    as_string_array(column).iter().map(|x| x.map(|x| x.as_bytes())),
+                    opts,
+                ),
+                DataType::LargeUtf8 => variable::encode(
+                    out,
+                    as_largestring_array(column)
+                        .iter()
+                        .map(|x| x.map(|x| x.as_bytes())),
+                    opts,
+                ),
+                _ => unreachable!(),
+            }
         }
-        DataType::LargeBinary => {
-            variable::encode(out, as_generic_binary_array::<i64>(column).iter(), opts)
+        Encoder::Dictionary(dict) => {
+            downcast_dictionary_array! {
+                column => encode_dictionary(out, column, dict, opts),
+                _ => unreachable!()
+            }
         }
-        DataType::Utf8 => variable::encode(
-            out,
-            as_string_array(column).iter().map(|x| x.map(|x| x.as_bytes())),
-            opts,
-        ),
-        DataType::LargeUtf8 => variable::encode(
-            out,
-            as_largestring_array(column)
-                .iter()
-                .map(|x| x.map(|x| x.as_bytes())),
-            opts,
-        ),
-        DataType::Dictionary(_, _) => downcast_dictionary_array! {
-            column => encode_dictionary(out, column, dictionary.unwrap(), opts),
-            _ => unreachable!()
+        Encoder::Struct(rows, null) => {
+            let array = as_struct_array(column.as_ref());
+            let null_sentinel = null_sentinel(opts);
+            out.offsets
+                .iter_mut()
+                .skip(1)
+                .enumerate()
+                .for_each(|(idx, offset)| {
+                    let (row, sentinel) = match array.is_valid(idx) {
+                        true => (rows.row(idx), 0x01),
+                        false => (*null, null_sentinel),
+                    };
+                    let end_offset = *offset + 1 + row.as_ref().len();
+                    out.buffer[*offset] = sentinel;
+                    out.buffer[*offset + 1..end_offset].copy_from_slice(row.as_ref());

Review Comment:
   > aren't accidentally compared with the contents of another field 
   
   The same way as normal, effectively this approach flattens the schema in a depth first manner. Lists will need encoding in this way, that's still to come :smile: 
   
   > Also I wonder why not encode the struct array directly into out rather than copy it from another Rows?
   
   Nulls, we need to not encode the values if the `StructArray` contains a null. Otherwise you would potentially establish an ordering between nulls, based on the "masked" values.



-- 
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: github-unsubscribe@arrow.apache.org

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