Re: [GENERAL] Constraint exclusion-like behavior for UNION ALL views

2017-03-29 Thread Tony Cebzanov
On 3/29/17 3:39 PM, David G. Johnston wrote: That said, I'm not sure what using materialized views instead of normal tables buys you in the first place. I could see possibly using a materialized view as the current month's table but the historical tables usually don't require refreshing. My e

Re: [GENERAL] Constraint exclusion-like behavior for UNION ALL views

2017-03-29 Thread David G. Johnston
On Wed, Mar 29, 2017 at 12:19 PM, Tony Cebzanov wrote: > Are either of these things that could be supported in the future? If not, > is there a better way to get this kind of behavior so that materialized > views are more useful when the amount of data increases and it's not > feasible to update

[GENERAL] Constraint exclusion-like behavior for UNION ALL views

2017-03-29 Thread Tony Cebzanov
PostgreSQL's materialized view functionality is very useful, but one problem is that when the view gets large, there is no way to refresh part of it. I know that table partitioning is coming in Postgres 10, but I haven't heard anything about ideas for partitioning / sharding of materialized vi