Hi,
On 2022-06-15 18:24:45 -0700, Mark Dilger wrote:
> > On Jun 15, 2022, at 6:01 PM, Andres Freund <andres@anarazel.de> wrote:
> > On 2022-06-15 17:21:56 -0700, Mark Dilger wrote:
> >> While developing various Table Access Methods, I have wanted a callback for
> >> determining if CLUSTER (and VACUUM FULL) should be run against a table
> >> backed by a given TAM. The current API contains a callback for doing the
> >> guts of the cluster, but by that time, it's a bit too late to cleanly back
> >> out. For single relation cluster commands, raising an error from that
> >> callback is probably not too bad. For multi-relation cluster commands, that
> >> aborts the clustering of other yet to be processed relations, which doesn't
> >> seem acceptable.
> >
> > Why not? What else do you want to do in that case? Silently ignoring
> > non-clusterable tables doesn't seem right either. What's the use-case for
> > swallowing the error?
>
> Imagine you develop a TAM for which the concept of "clustering" doesn't have
> any defined meaning. Perhaps you've arranged the data in a way that has no
> similarity to heap, and now somebody runs a CLUSTER command (with no
> arguments.)
I think nothing would happen in this case - only pre-clustered tables get
clustered in an argumentless CLUSTER. What am I missing?
Greetings,
Andres Freund
