Btree bloat query changelog - part 3

Here are some fresh news about my previous work (see part 1 and part 2) on a better Btree bloat query.

Changelog

• add field is_na to filter out indexes we can not estimate the bloat with “accuracy” (currently only indexes referencing a fields using the name type).
• better estimation for variable length types (varchar, text, bytea, …). Thequery now try to consider their header part
• Three different queries depending on the PostgreSQL version:
  • from 7.4 to 8.1: https://gist.github.com/ioguix/5f60e24a77828078ff5f
  • for 8.2: https://gist.github.com/ioguix/0675875e2780b362ef28
  • for 8.3 and more: https://gist.github.com/ioguix/c29d5790b8b93bf81c27
• ignore non-valid indexes for 8.2 and more

Known issue

While working on table bloat (I will blog about that very soon), I found a large deviation on statistics on array types. I’m not sure how to handle correctly these animals’ header yet.

Cheers and happy monitoring!