Support for partial/conditional indexes

Searching about it looks like there is a JIRA for supporting partial / conditional indexes dating back to 2009 and some occasional inquiries on the subject. Is there a particular reason why support for these has languished for so long? Is there just very little demand for the feature or are there significant technical hurdles in adding this functionality? We make extensive use of partial indexes in our user of Postgres (v10) and support for these both in changelogs and in the tooling for changelog generation from existing schema would be valuable to us. I would prefer not to have to fall-back on calling out from the JVM to pg_dump and pg_restore for schema migration. Thanks for any insights you can provide.

Hi @jamey_clari,

Sorry for the delay in responding. I am including folks from our product team (@mariochampion and @Pete). The use case sounds interesting and could answer better on the level of support currently and future for partial / conditional indexes.



Thanks Ronak, it’d be great to see some support for this stuff as usage has to be fairly common as DB sizes grow and indexing becomes more challenging.