getChan commented on code in PR #20432: URL: https://github.com/apache/datafusion/pull/20432#discussion_r2834131229
########## datafusion/sqllogictest/test_files/projection_pushdown.slt: ########## @@ -1949,3 +1949,27 @@ ORDER BY simple_struct.id; 3 3 4 4 5 5 + +# ========================================================================= +# Regression: user-provided __datafusion_extracted aliases must not +# collide with optimizer-generated ones +# (https://github.com/apache/datafusion/issues/20430) +# ========================================================================= + +statement ok +COPY ( select {f1: 1, f2: 2} as s +) TO 'test_files/scratch/projection_pushdown/test.parquet' +STORED AS PARQUET; + +statement ok +CREATE EXTERNAL TABLE t +STORED AS PARQUET +LOCATION 'test_files/scratch/projection_pushdown/test.parquet'; + +query I +SELECT + get_field(s, 'f1') AS __datafusion_extracted_2 +FROM t +WHERE COALESCE(get_field(s, 'f1'), get_field(s, 'f2')) = 1; +---- +1 Review Comment: Would it make sense to also add an `EXPLAIN` assertion here? It could guard the alias allocation behavior directly (e.g. user-provided `__datafusion_extracted_2` remains stable while optimizer-generated aliases move to the next IDs), so future optimizer refactors don’t regress silently. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
