You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Comet native writer works for basic ops, however to move forward and keep compatibility with Spark we need to run Spark writer and partition discovery tests with Comet writer enabled first to make sure we have a solid reliable base for Comet native writer changes
Common across all Spark versions (3.4, 3.5, 4.0) :
1. empty file should be skipped while write to file
2. INSERT INTO TABLE - complex type but different names
3. Insert overwrite table command should output correct schema: basic
4. parquet timestamp conversion
5. SPARK-29174 Support LOCAL in INSERT OVERWRITE DIRECTORY to data source
6. SPARK-33901: ctas should should not change table's schema
7. SPARK-37160: CREATE TABLE AS SELECT with CHAR_AS_VARCHAR
8. SPARK-38336 INSERT INTO statements with tables with default columns: positive tests
9. SPARK-38811 INSERT INTO on columns added with ALTER TABLE ADD COLUMNS: Positive tests
10. SPARK-43071: INSERT INTO from queries whose final operators are not projections
11. write path implements onTaskCommit API correctly
12. Write Spark version into Parquet metadata
Additional tests which had to be ignored for spark 4 :
1. ctas with union
2. SPARK-48817: test multi inserts
Total: 12 tests (Spark 3.4/3.5) | 14 tests (Spark 4.0)
Activity
feat: add filesystem utilities for testing write commit protocol
test: add comprehensive 2PC tests for native Parquet writer
coderfender commentedon Jan 25, 2026
Looking into this
[-]Run Spark tests with Comet native runner[/-][+]Run Spark tests with Comet native writer[/+]coderfender commentedon Jan 30, 2026
Updated Spark 4 diff file (ignoring tests) .
coderfender commentedon Feb 1, 2026
take
coderfender commentedon Feb 3, 2026
Ignored Tests for Comet Native Parquet Writer