test: split write.rs into topic-focused files (delta-io#2460)

## What changes are proposed in this pull request?

Splits the ~4800-line `kernel/tests/write.rs` into 13 topic-focused
files plus a shared `kernel/tests/common/write_utils.rs` for helpers.


Closes delta-io#2458.

## How was this change tested?

Author: DrakeLin

kernel/src/transaction/mod.rs

@@ -2111,7 +2111,7 @@ mod tests {
 
     // Note: Additional test coverage for partial file matching (where some files in a scan
     // have DV updates but others don't) is provided by the end-to-end integration test
-    // kernel/tests/dv.rs and kernel/tests/write.rs, which exercises
+    // kernel/tests/dv.rs and kernel/tests/write_remove_dv.rs, which exercise
     // the full deletion vector write workflow including the DvMatchVisitor logic.
 
     #[test]

kernel/tests/README.md

@@ -9,7 +9,7 @@ Test tables organized by feature area. Tables live in two locations:
 
 | Table | Location | Schema | Protocol (R/W) | Features | Description | Tests |
 |-------|----------|--------|----------|----------|-------------|-------|
-| `table-with-dv-small` | data/ | `value: int` | v3/v7 | r:`deletionVectors` w:`deletionVectors` | 10 rows, 2 soft-deleted by DV, 8 visible. Most heavily referenced test table. | `dv.rs::test_table_scan(with_dv)`, `write.rs::test_remove_files_adds_expected_entries`, `write.rs::test_update_deletion_vectors_adds_expected_entries`, `read.rs::with_predicate_and_removes`, `path.rs::test_to_uri/test_child/test_child_escapes`, `snapshot.rs::test_snapshot_read_metadata/test_new_snapshot/test_snapshot_new_from/test_read_table_with_missing_last_checkpoint/test_log_compaction_writer`, `deletion_vector.rs` tests, `transaction/mod.rs::setup_dv_enabled_table/test_add_files_schema/test_new_deletion_vector_path`, `default/parquet.rs` read test, `default/json.rs` read test, `log_compaction/tests.rs::create_mock_snapshot`, `resolve_dvs.rs` tests |
+| `table-with-dv-small` | data/ | `value: int` | v3/v7 | r:`deletionVectors` w:`deletionVectors` | 10 rows, 2 soft-deleted by DV, 8 visible. Most heavily referenced test table. | `dv.rs::test_table_scan(with_dv)`, `write_remove_dv.rs::test_remove_files_adds_expected_entries`, `write_remove_dv.rs::test_update_deletion_vectors_adds_expected_entries`, `read.rs::with_predicate_and_removes`, `path.rs::test_to_uri/test_child/test_child_escapes`, `snapshot.rs::test_snapshot_read_metadata/test_new_snapshot/test_snapshot_new_from/test_read_table_with_missing_last_checkpoint/test_log_compaction_writer`, `deletion_vector.rs` tests, `transaction/mod.rs::setup_dv_enabled_table/test_add_files_schema/test_new_deletion_vector_path`, `default/parquet.rs` read test, `default/json.rs` read test, `log_compaction/tests.rs::create_mock_snapshot`, `resolve_dvs.rs` tests |
 | `table-without-dv-small` | data/ | `value: long` | v1/v2 | | 10 rows, all visible. Companion to table-with-dv-small. | `dv.rs::test_table_scan(without_dv)`, `transaction/mod.rs::setup_non_dv_table/create_existing_table_txn/test_commit_io_error_returns_retryable_transaction`, `sequential_phase.rs::test_sequential_v2_with_commits_only/test_sequential_finish_before_exhaustion_error`, `parallel_phase.rs` tests, `scan/tests.rs::test_scan_metadata_paths/test_scan_metadata/test_scan_metadata_from_same_version` |
 | `with-short-dv` | data/ | `id: long, value: string, timestamp: timestamp, rand: double` | v3/v7 | r:`deletionVectors` w:`deletionVectors` | 2 files x 5 rows. First file has inline DV (`storageType="u"`) deleting 3 rows. | `read.rs::short_dv` |
 | `dv-partitioned-with-checkpoint` | golden_data/ | `value: int, part: int` partitioned by `part` | v3/v7 | r:`deletionVectors` w:`deletionVectors` | DVs on a partitioned table with a checkpoint | `golden_tables.rs::golden_test!` |
@@ -41,9 +41,9 @@ Test tables organized by feature area. Tables live in two locations:
 
 | Table | Location | Schema | Protocol (R/W) | Features | Description | Tests |
 |-------|----------|--------|----------|----------|-------------|-------|
-| `partition_cm/none` | data/ | `value: int, category: string` partitioned by `category` | v1/v1 | `columnMapping.mode=none` | Partitioned write with CM disabled | `write.rs::test_column_mapping_partitioned_write(cm_none)` |
-| `partition_cm/id` | data/ | `value: int, category: string` partitioned by `category` | v3/v7 | r:`columnMapping` w:`columnMapping`, `columnMapping.mode=id` | Partitioned write with CM id mode | `write.rs::test_column_mapping_partitioned_write(cm_id)` |
-| `partition_cm/name` | data/ | `value: int, category: string` partitioned by `category` | v3/v7 | r:`columnMapping` w:`columnMapping`, `columnMapping.mode=name` | Partitioned write with CM name mode | `write.rs::test_column_mapping_partitioned_write(cm_name)` |
+| `partition_cm/none` | data/ | `value: int, category: string` partitioned by `category` | v1/v1 | `columnMapping.mode=none` | Partitioned write with CM disabled | `write_column_mapping.rs::test_column_mapping_partitioned_write(cm_none)` |
+| `partition_cm/id` | data/ | `value: int, category: string` partitioned by `category` | v3/v7 | r:`columnMapping` w:`columnMapping`, `columnMapping.mode=id` | Partitioned write with CM id mode | `write_column_mapping.rs::test_column_mapping_partitioned_write(cm_id)` |
+| `partition_cm/name` | data/ | `value: int, category: string` partitioned by `category` | v3/v7 | r:`columnMapping` w:`columnMapping`, `columnMapping.mode=name` | Partitioned write with CM name mode | `write_column_mapping.rs::test_column_mapping_partitioned_write(cm_name)` |
 | `table-with-columnmapping-mode-name` | golden_data/ | `ByteType: byte, ShortType: short, IntegerType: int, LongType: long, FloatType: float, DoubleType: double, decimal: decimal(10,2), BooleanType: boolean, StringType: string, BinaryType: binary, DateType: date, TimestampType: timestamp, nested_struct: struct{aa: string, ac: struct{aca: int}}, array_of_prims: array<int>, array_of_arrays: array<array<int>>, array_of_structs: array<struct{ab: long}>, map_of_prims: map<int,long>, map_of_rows: map<int,struct{ab: long}>, map_of_arrays: map<long,array<int>>` | v2/v5 | `columnMapping.mode=name` | Column mapping name mode | `golden_tables.rs::golden_test!` |
 | `table-with-columnmapping-mode-id` | golden_data/ | `ByteType: byte, ShortType: short, IntegerType: int, LongType: long, FloatType: float, DoubleType: double, decimal: decimal(10,2), BooleanType: boolean, StringType: string, BinaryType: binary, DateType: date, TimestampType: timestamp, nested_struct: struct{aa: string, ac: struct{aca: int}}, array_of_prims: array<int>, array_of_arrays: array<array<int>>, array_of_structs: array<struct{ab: long}>, map_of_prims: map<int,long>, map_of_rows: map<int,struct{ab: long}>, map_of_arrays: map<long,array<int>>` | v2/v5 | `columnMapping.mode=id` | Column mapping id mode | `golden_tables.rs::golden_test!` |
 
@@ -52,7 +52,7 @@ Test tables organized by feature area. Tables live in two locations:
 | Table | Location | Schema | Protocol (R/W) | Features | Description | Tests |
 |-------|----------|--------|----------|----------|-------------|-------|
 | `with_checkpoint_no_last_checkpoint` | data/ | `letter: string, int: long, date: date` | v1/v2 | `checkpointInterval=2` | Checkpoint at v2 but missing `_last_checkpoint` hint file | `snapshot.rs::test_read_table_with_checkpoint`, `scan/tests.rs::test_scan_with_checkpoint`, `sequential_phase.rs::test_sequential_checkpoint_no_commits`, `checkpoint_manifest.rs` tests, `sync/parquet.rs` test, `default/parquet.rs` test |
-| `external-table-different-nullability` | data/ | `i: int` | v1/v2 | `checkpointInterval=2` | Parquet files have different nullability than Delta schema; includes checkpoint | `write.rs::test_checkpoint_non_kernel_written_table` |
+| `external-table-different-nullability` | data/ | `i: int` | v1/v2 | `checkpointInterval=2` | Parquet files have different nullability than Delta schema; includes checkpoint | `write_clustered.rs::test_checkpoint_non_kernel_written_table` |
 | `checkpoint` | golden_data/ | `intCol: int` | v1/v2 | | Basic checkpoint read | `golden_tables.rs::golden_test!(checkpoint_test)` |
 | `corrupted-last-checkpoint-kernel` | golden_data/ | `id: long` | v1/v2 | | Corrupted `_last_checkpoint` file | `golden_tables.rs::golden_test!` |
 | `multi-part-checkpoint` | golden_data/ | `id: long` | v1/v2 | `checkpointInterval=1` | Multi-part checkpoint files | `golden_tables.rs::golden_test!` |

kernel/tests/common/mod.rs

@@ -1,3 +1,5 @@
+pub mod write_utils;
+
 #[macro_export]
 macro_rules! sort_lines {
     ($lines: expr) => {{