"show stats_xxx" is slow if there are massive tables

## Enhancement

This is a SaaS user with more than 190,000 tables. `show stats_xxx` is too slow in this scenario. See the example below, `show stats_buckets where db_name=? and table_name=? and column_name=?` takes around 2mins to return the result:

Start by locating the implementation of the `show stats_buckets` command and reproduce the reported query with a database containing more than 190,000 tables. Trace why filtering by `db_name`, `table_name`, and `column_name` takes about two minutes; done means the command returns substantially faster for this scenario, with regression coverage where the relevant tests are found.

Written by the indexing model from the issue text.