[GitHub] [spark] grundprinzip commented on pull request #38468: [SPARK-41005][CONNECT][PYTHON] Arrow-based collect

2022-11-10 Thread GitBox
grundprinzip commented on PR #38468: URL: https://github.com/apache/spark/pull/38468#issuecomment-1310307256 I would like to close the discussion on the ordered vs un-ordered result. 1) For simple clients ordered results are what they expect and it follows the precedent of what users

[GitHub] [spark] grundprinzip commented on pull request #38468: [SPARK-41005][CONNECT][PYTHON] Arrow-based collect

2022-11-03 Thread GitBox
grundprinzip commented on PR #38468: URL: https://github.com/apache/spark/pull/38468#issuecomment-1303010466 @zhengruifeng can you please add test cases for things like `select * from table limit 0` where the optimizer decides there are no qualifying rows but we have to return an empt