On 6/1/20 4:58 AM, Peter J. Holzer wrote: [snip]
As a developer (and part time DBA) I have a hard time thinking of any Oracle feature that I'm missing in PostgreSQL.
The inability to do a point-in-time restoration of a *single* database in a multi-db cluster is a serious -- and fundamental -- missing feature (never to be implemented because of the fundamental design).
In SQL Server, it's trivial to restore -- including differentials and WAL files -- an old copy of a prod database *to a different name* so that you now have databases FOO and FOO_OLD in the same instance.
In Postgres, though, you've got to create a new cluster using a new port number (which in our case means sending a firewall request through channels and waiting two weeks while the RISK team approves opening the port -- and they might decline it because it's non-standard -- and then the Network team creates a /change order/ and then implements it).
Bottom line: something I can do in an afternoon with SQL Server takes two weeks for Postgres.
This has given Postgres a big, fat black eye with our end users. -- Angular momentum makes the world go 'round.
