This is an automated email from the ASF dual-hosted git repository.
Jefffrey pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/arrow-rs.git
The following commit(s) were added to refs/heads/main by this push:
new 852d4a9bc2 Replace BufferBuilder with Vec in MapArrayDecoder (#10629)
852d4a9bc2 is described below
commit 852d4a9bc27a33e256b802cac46e828e5d1a9915
Author: cakeni <[email protected]>
AuthorDate: Thu Aug 20 09:23:34 2026 +0800
Replace BufferBuilder with Vec in MapArrayDecoder (#10629)
# Which issue does this PR close?
- Part of #10245.
# Rationale for this change
Using `Vec` instead of `BufferBuilder` can improve performance by
benefiting from Rust's optimized `Vec` implementation. This updates the
remaining `BufferBuilder` usage in `arrow-json` and follows the existing
`ListArrayDecoder` approach.
# What changes are included in this PR?
- Replace `MapArrayDecoder`'s offset `BufferBuilder<i32>` with a
preallocated `Vec<i32>`.
- Construct the final `ScalarBuffer` directly from the vector.
- Remove the now-unused `BufferBuilder` import.
# Are these changes tested?
Yes. The following checks pass:
- `cargo +stable-x86_64-pc-windows-gnu fmt --all -- --check`
- `cargo +stable-x86_64-pc-windows-gnu clippy -p arrow-json
--all-targets --all-features --no-deps -- -D warnings`
- `cargo +stable-x86_64-pc-windows-gnu test -p arrow-json
--all-features` (134 unit tests and 17 doctests passed)
# Are there any user-facing changes?
No.
## AI assistance
OpenAI Codex assisted with code exploration and drafting this change. I
reviewed the implementation and final diff.
Co-authored-by: Jeffrey Vo <[email protected]>
---
arrow-json/src/reader/map_array.rs | 9 ++++-----
1 file changed, 4 insertions(+), 5 deletions(-)
diff --git a/arrow-json/src/reader/map_array.rs
b/arrow-json/src/reader/map_array.rs
index abb035193c..cd8f868e31 100644
--- a/arrow-json/src/reader/map_array.rs
+++ b/arrow-json/src/reader/map_array.rs
@@ -17,7 +17,6 @@
use std::sync::Arc;
-use arrow_array::builder::BufferBuilder;
use arrow_array::{ArrayRef, MapArray, StructArray};
use arrow_buffer::{ArrowNativeType, NullBufferBuilder, OffsetBuffer,
ScalarBuffer};
use arrow_schema::{ArrowError, DataType, FieldRef, Fields};
@@ -83,8 +82,8 @@ impl MapArrayDecoder {
impl ArrayDecoder for MapArrayDecoder {
fn decode(&mut self, tape: &Tape<'_>, pos: &[u32]) -> Result<ArrayRef,
ArrowError> {
- let mut offsets = BufferBuilder::<i32>::new(pos.len() + 1);
- offsets.append(0);
+ let mut offsets = Vec::with_capacity(pos.len() + 1);
+ offsets.push(0);
let mut key_pos = Vec::with_capacity(pos.len());
let mut value_pos = Vec::with_capacity(pos.len());
@@ -122,7 +121,7 @@ impl ArrayDecoder for MapArrayDecoder {
let offset = i32::from_usize(key_pos.len()).ok_or_else(|| {
ArrowError::JsonError("offset overflow decoding
MapArray".to_string())
})?;
- offsets.append(offset)
+ offsets.push(offset)
}
assert_eq!(key_pos.len(), value_pos.len());
@@ -139,7 +138,7 @@ impl ArrayDecoder for MapArrayDecoder {
let nulls = nulls.as_mut().and_then(|x| x.finish());
// SAFETY: offsets are built monotonically starting from 0
- let offsets = unsafe {
OffsetBuffer::new_unchecked(ScalarBuffer::from(offsets.finish())) };
+ let offsets = unsafe {
OffsetBuffer::new_unchecked(ScalarBuffer::from(offsets)) };
let array = MapArray::try_new(
self.entries_field.clone(),
