That was a bit of what we discussed at the sync this morning. Whether we should have a generic enable feature update for one way features that we don’t have the ability to disable. We couldn’t come up with more examples of features we actually wanted to add. I think I’d we have at least two more it would make sense.
On Wed, Jan 22, 2025 at 6:04 PM Daniel Weeks <[email protected]> wrote: > Just a minor question added to the PR. We're adding an explicit 'enable' > as an update type and I wonder if it would be better to generalize it so > that we don't have separate updates to disable/enable (more forward > thinking as this is the first case quite like this). > > -Dan > > > > On Wed, Jan 22, 2025 at 3:55 PM Amogh Jahagirdar <[email protected]> wrote: > >> +1 Thanks Russell >> >> On Wed, Jan 22, 2025 at 4:50 PM [email protected] <[email protected]> >> wrote: >> >>> +1 >>> >>> On Wed, Jan 22, 2025 at 2:51 PM Russell Spitzer < >>> [email protected]> wrote: >>> >>>> Hey Y'all >>>> >>>> Yet another Row Lineage Spec update. This adds a MetadataUpdate >>>> EnableRowLineage to the REST Spec. We briefly talked today >>>> about an alternative EnableFeature(Feature Name) API instead but in the >>>> absence of other features it doesn't seem >>>> like that's really a requirement now. >>>> >>>> I agreed that if we ever do have another feature we want to enable in a >>>> similar way I would take the blame for adding >>>> this API rather than a generic one. >>>> >>>> That said please take a look >>>> >>>> https://github.com/apache/iceberg/pull/12050 >>>> >>>> >>>> Note: We only allow enabling row lineage, it cannot be disabled. >>>> >>>> Thanks for your time, >>>> Russ >>>> >>>
