The intention is to set an upper limit for table size while keeping as much
snapshots as possible. Setting max-snapshot-age-ms to a small value will
lose history for some tables while setting min-snapshots-to-keep to a
medium value will keep too much history for others.
Lewis, William 于2025年1月18日
To clarify, the intention is that if max-snapshots-to-keep is set, then
snapshots will be expired even if they are younger than max-snapshot-age-ms?
Does this solve a problem that isn’t solved by setting max-snapshot-age-ms to a
small value and setting min-snapshots-to-keep to your desired value
+0, as I agree with Amogh, I think it would fit nicely with Honah's work of
formalizing the properties.
Kind regards,
Fokko
Op vr 17 jan 2025 om 08:55 schreef Honah J. :
> +1
>
> Best,
> Honah
>
> On Thu, Jan 16, 2025 at 22:54 Manish Malhotra <
> manish.malhotra.w...@gmail.com> wrote:
>
>> +1,
Thanks everyone, especially Amogh and Fokko for providing other views on
this change. I think your viewpoints are definitely valid although I think
currently we should be working on moving things away from Snapshot Summary
if at all possible. I think it's better if the information there is never
cr
+1 (non binding)
Regards
JB
On Wed, Jan 15, 2025 at 5:59 PM Russell Spitzer
wrote:
>
> Hi Everyone!
>
> PR: https://github.com/apache/iceberg/pull/11976/files
>
> Split out from #11948
>
> Working on the row-lineage implementation made it clear that we needed a way
> to get information from the
