ayushtkn commented on PR #6394: URL: https://github.com/apache/hive/pull/6394#issuecomment-4164190174
> Does it mean that in case of SerializedTable we always load a fresh copy, Yep, in case we don't have that in cache, if we have that in cache, we always attempt to find it in cache. I think it would be very few cases else the code would have blasted entirely, which it didn't. or our test coverage has some issues. The case where it isn't present in cache is when you aren't going via the normal sql route like in some tests, else postGetTable hook would always load the Iceberg Table if my memory treats me well, to fetch the format version from the metadata.json, So, for sql queries it should be there for sure. > seems inefficient as well. It is for those cases for sure, but we don't introduce that over here, so kind of something we can figure out, if we find some way or if it actually hits us so hard. > why serialize/deserialize at all? I honestly don't have context around this, that logic got grandfathered to us via the original iceberg repo & I wasn't involved then but usually these jobConf things are put if it has to be used in some other context outside the original session, there are cases like that. I know during write it is there for sure, read side I am not sure -- 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]
