openinx opened a new issue #2308: URL: https://github.com/apache/iceberg/issues/2308
Since we've introduced row-level delete in format v2, so we will encounter the problem that RewriteFiles and RowDelta commit the transaction at the same time. Assume that we have an iceberg table `test`, and has the following events: ```python INSERT <1, 'AAA'> INSERT <2, 'BBB'> DELETE <1, 'AAA'> ``` At the timestamp `t1`, someone start a rewrite action to rewrite the whole table. At the timestamp `t2`, someone start another transaction to update the rows in table `test`: ```python DELETE <2, 'BBB'> ``` At the timestamp `t3`, the update txn (which started from `t2`) commit the txn successfully. At the timestamp `t1`, the rewrite action commit the txn successfully. Finally, the table will have one row `<2, 'BBB'>`, while in fact we should have no rows. That's an unexpected bug after introducing format v2, and we will need solution to handle it. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
