2 ms·
It seems like the inverse of a merge table. Years ago we used to archive data into tables named by the date range and then we would create a merge table that c
by rootusrootus 9y ago
It seems like the inverse of a merge table. Years ago we used to archive data into tables named by the date range and then we would create a merge table that combined them. So every night we would drop the oldest, create a new one, and create a new merge table of the updated set.
Partitioning looks to be going the other way. Create a big table, partition for convenience, drop partitions easily when you are ready to dispose of them. I can definitely see some cases where this would be super handy.
- pgaddict 9y agoI'm not familiar with merge tables, but it sounds quite like a view for UNION of all the tables.