Balázs Németh
Balázs Németh
I have no time to do it properly with docs, tests and stuff nowadays, but here is the pure code itself for anyone willing to finish it: ```java import java.lang.reflect.Type;...
slighty related https://github.com/apache/beam/issues/22631
@joey-berman I'm on 2.40.0 now and did not face this issue recently, but I have no idea if it's just coincidence, or it's a legit fix.
@joey-berman I did not, but I have non-zero getAllowedTimestampSkew on my DoFns, so if I get the message right this can't happen to me anyway.
So basically the ability to support a provider that calculates an offset to be used if it's not found, and the default provider being the current implementation that throws an...
I'm not sure I will have time in the near future to implement this, but given how slowly the discussion went I created a working solution for reading the whole...
@edwardmedia I did, and I get why is it like that, BUT I don't see how is that relevant on the long term though. IMO it's not a limitation that...
@edwardmedia I have updated the issue with original + modified config, and the received output.
Well I think you missed probably the most important business aspect of this issue. The table should never be dropped, and that deletion protection should never be turned off. I...
@edwardmedia Using tables.patch instead of table.update would change the "replace" behaviour? I don't see how it's coming from the API at all, so why would that change it? "this is...