Re: [VOTE] Deletion Vectors in V3

2024-10-29 Thread Jean-Baptiste Onofré
+1 (non binding) Regards JB On Tue, Oct 29, 2024 at 10:45 PM Anton Okolnychyi wrote: > > Hi folks, > > We have been discussing the new layout for position deletes in V3 for a while > now. It seems the community reached consensus. I'd like to start a vote on > adding deletion vectors to the V3

Re: [VOTE] Deletion Vectors in V3

2024-10-29 Thread Gang Wu
+1 (non-binding) Best, Gang On Wed, Oct 30, 2024 at 5:46 AM Anton Okolnychyi wrote: > Hi folks, > > We have been discussing the new layout for position deletes in V3 for a > while now. It seems the community reached consensus. I'd like to start a > vote on adding deletion vectors to the V3 spec

[VOTE] Deletion Vectors in V3

2024-10-29 Thread Anton Okolnychyi
Hi folks, We have been discussing the new layout for position deletes in V3 for a while now. It seems the community reached consensus. I'd like to start a vote on adding deletion vectors to the V3 spec as described in PRs 11238 and 11240

Re: [DISCUSS] Discrepancy Between Iceberg Spec and Java Implementation for Snapshot summary's 'operation' key

2024-10-29 Thread Russell Spitzer
I think that "read bad, write good" is an ok direction forward. The warnings are a good touch as well, we want to make sure that if someone is building an implementation of Iceberg and they test against the reference implementation, it will be clear whether they are to spec or not. On Tue, Oct 29

Re: [DISCUSS] Discrepancy Between Iceberg Spec and Java Implementation for Snapshot summary's 'operation' key

2024-10-29 Thread Fokko Driesprong
I think we more or less agree here, and as Russel points out, it is a bit of a chicken and egg. I would suggest reading the faulty metadata and emitting a warning that it defaults to overwrite: - Java: https://github.com/apache/iceberg/pull/11421 - PyIceberg: https://github.com/apache/iceber

Re: [DISCUSS] Partial Metadata Loading

2024-10-29 Thread Gabor Kaszab
Hi Iceberg Community, I just wanted to mention that I was also going to start a discussion about getting partial information from LoadTableResponse through the REST API. My motivation is a bit different here, though: Impala currently has strong integration with HMS and in turn with the HiveCatalog