tkhurana commented on PR #2611: URL: https://github.com/apache/phoenix/pull/2611#issuecomment-5668103654
Some test gaps identified - (A) SINGLE_CELL storage-scheme partial upsert is entirely untested. Both new GlobalIndexCheckerIT tests early-return on encoded and hardcode ONE_CELL_PER_COLUMN — and encoded=true (→ SINGLE_CELL_ARRAY_WITH_OFFSETS) is exactly the matrix combination they skip. The single-cell branch of getAllColumnsForDataTable/addColumnRefForScan is the PR's own load-bearing fast-path claim and has zero assertion behind it. This is the one gap I'd most want closed (matching data+index single-cell scheme, full then partial upsert, assert index/data agreement). - (B) Index-WHERE (partial index) columns — getAllColumnsForDataTable deliberately includes them, but no test creates CREATE INDEX … WHERE … on an immutable table and omits the WHERE column. - (C) Delete over an immutable ROW_TIMESTAMP table with an index — the exact branch the DeleteCompiler signature change was added for is unexercised (the one DeleteIT change is about single-row plans, not ROW_TIMESTAMP). - (D) Multiple global indexes on one table — the cross-maintainer column union in isPartialGlobalIndexMutation is untested. -- 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]
