Em dom., 21 de abr. de 2024 às 22:35, David Rowley <dgrowleyml@gmail.com> escreveu:
Both VACUUM FULL and CLUSTER go through a very similar code path. Both use cluster_rel(). VACUUM FULL just won't make use of an existing index to provide presorted input or perform a sort, whereas CLUSTER will attempt to choose the cheapest out of these two to get sorted results.
But wouldn't it be good that VACUUM FULL uses that index defined by Cluster, if it exists ? Maybe an additional option for VACUUM FULL ?
Because if I periodically reorganize all tables I have to run CLUSTER once, which will run on clustered tables, and VACUUM FULL on every table that is not clustered, because if I run VACUUM FULL for entire database it'll just ignore cluster index defined for every table. So, INDISCLUSTERED is used when running CLUSTER but is ignored when running VACUUM FULL.