refactor(streaming): use ChangeBuffer for executors

[BugenZhao]

I hereby agree to the terms of the RisingWave Labs, Inc. Contributor License Agreement.

What's changed and what's your intention?

With the introduction of the ChangeBuffer in the recent PR #23507, we can now streamline the implementation of executors that previously replicated this functionality, including:

• MaterializeCache
• TopNStaging
• OverWindowExecutor

Actually, this was once proposed in #19451 (comment), where it seemed that we merged the PR in a rush and didn't get it done eventually.

Checklist

• I have written necessary rustdoc comments.
• I have added necessary unit tests and integration tests.
• I have added test labels as necessary.
• I have added fuzzing tests or opened an issue to track them.
• My PR contains breaking changes.
• My PR changes performance-critical code, so I will run (micro) benchmarks and present the results.
• I have checked the Release Timeline and Currently Supported Versions to determine which release branches I need to cherry-pick this PR into.

Documentation

• My PR needs documentation updates.

Release note

[BugenZhao]

• refactor(streaming): use ChangeBuffer for executors #23508 👈 (View in Graphite)
• refactor(streaming): extract ChangeBuffer for chunk compaction #23507
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[BugenZhao]

This doesn't matter any more as we switch internal to IndexMap.

The caller guarantees that when updating a row in staging, it always call delete before insert. Thus, by iterating the changes based on the operation order, we won't violate the LIMIT either.

[Copilot]

Pull Request Overview

This PR refactors streaming executors to use the centralized ChangeBuffer utility instead of maintaining duplicated change-buffering logic. The change eliminates redundant implementations in MaterializeCache, TopNStaging, and OverWindowExecutor, improving code maintainability and consistency.

Key Changes:

• Replaced custom change buffering logic with ChangeBuffer in multiple executors
• Added try_map and map methods to Record<R> for row transformation
• Updated executors to use Record types and append_record instead of tuples

Reviewed Changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
src/stream/src/common/change_buffer.rs	Added Default trait implementation for ChangeBuffer
src/common/src/array/stream_record.rs	Added try_map and map methods to Record<R> for row transformations
src/stream/src/executor/mview/materialize.rs	Removed local ChangeBuffer implementation in favor of common ChangeBuffer; removed generate_output function
src/stream/src/executor/over_window/general.rs	Replaced manual change merging logic with ChangeBuffer
src/stream/src/executor/top_n/top_n_cache.rs	Refactored TopNStaging to use ChangeBuffer internally; simplified insert/delete operations
src/stream/src/executor/top_n/top_n_plain.rs	Updated to use Record type and append_record method
src/stream/src/executor/top_n/top_n_appendonly.rs	Updated to use Record type and append_record method
src/stream/src/executor/top_n/group_top_n_appendonly.rs	Added explicit type annotation for stagings map; updated to use append_record
src/stream/src/executor/top_n/group_top_n.rs	Added explicit type annotation for stagings map; updated to use append_record

[chenzl25]

LGTM