vacuumdb: Teach vacuum_one_database() to reuse query results.
Presently, each call to vacuum_one_database() queries the catalogs
to retrieve the list of tables to process. A follow-up commit will
add a "missing stats only" feature to --analyze-in-stages, which
requires saving the catalog query results (since tables without
statistics will have them after the first stage). This commit adds
a new parameter to vacuum_one_database() that specifies either a
previously-retrieved list or a place to return the catalog query
results. Note that nothing uses this new parameter yet.
Author: Corey Huinker <
[email protected]>
Co-authored-by: Nathan Bossart <[email protected]>
Reviewed-by: John Naylor <[email protected]>
Discussion: https://postgr.es/m/Z5O1bpcwDrMgyrYy%40nathan