4 ms·
Native Table Partitioning in PostgreSQL 10
- comboy 10y agoduplicate: https://news.ycombinator.com/item?id=13128732 https://news.ycombinator.com/item?id=13128732 also I think the link and title of this post don't match
- jadbox 10y agoAnyone willing explain this to someone without context?
- Zikes 10y agoTable partitions allow for automatic sharding of data across multiple tables, while treating them as a single collection via a parent table.
- agotterer 10y agoWhat is the benefit over PG schemas?
- qaq 10y agoWhy are you comparing partitioned tables with schemas?
- grzm 10y agoTable partitions are for managing data that would ideally belong in the same table that for management reasons you'd like to split into multiple tables. One application is for managing time series data where you only want to keep the most recent data. For example, you want to save the most recent three months. For implementation reasons, running DELETE FROM big_table WHERE some_timestamp < '2016-09-01' is going to be painful. If you partition big_table by month, you can instead issue 'DROP big_table_201609' when you're done with it. Much cleaner and faster. You could do this in previous versions, through convention and scripts. Having it built into PostgreSQL should make this a lot easier. Schemas are something else entirely.
- jimktrains2 10y agoDid you mean to link to https://git.postgresql.org/gitweb/?p=postgresql.git;a=commit;h=f0e44751d7175fa3394da2c8f85e3ceb3cdbfe63 https://git.postgresql.org/gitweb/?p=postgresql.git;a=commit... This is a commit about synchronous replication not table partitioning.
- craigkerstiens 10y agoDoh, yes, for some reason wrong link got posted. Bad copy and paste error.