mirageyjd commented on issue #8932: URL: https://github.com/apache/iceberg/issues/8932#issuecomment-1784723175
Hi @RussellSpitzer. I ran an experiment on a table with 100K data files and collected the total cost of invoking [addManifest](https://github.com/apache/iceberg/blob/b56ec5ed421cb82d0c87c4fdd2a579db40f0e036/core/src/main/java/org/apache/iceberg/BaseRewriteManifests.java#L141). With `compatibility.snapshot-id-inheritance.enabled=true`, the cost is 2ms. With `compatibility.snapshot-id-inheritance.enabled=false`, the total cost is 13315ms. When snapshot id inheritance is disabled, all added manifests need to be validated, and then copied to a new file without changing the content. Copying manifests is expensive since it is executed in a single thread. https://github.com/apache/iceberg/blob/b56ec5ed421cb82d0c87c4fdd2a579db40f0e036/core/src/main/java/org/apache/iceberg/BaseRewriteManifests.java#L151-L157 -- 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]
