> On Nov 21, 2019, at 9:35 AM, Marc Millas <marc.mil...@mokadb.com> wrote: > > Hi Laurenz, > > > I was writing select from ""table"" as a template. We have to do this for a > bunch of tables. > So, to my understanding, what you suggest is to PITR up to the first > timestamp, extract all meaningfull tables, and then pitr to the second > timestamp > so as to be able to script a kind of "diff" between the 2 to get what I want. > > Yes ? > > Marc MILLAS > Senior Architect > +33607850334 > www.mokadb.com <http://www.mokadb.com/> > > > > On Thu, Nov 21, 2019 at 5:16 PM Laurenz Albe <laurenz.a...@cybertec.at > <mailto:laurenz.a...@cybertec.at>> wrote: > On Thu, 2019-11-21 at 17:07 +0100, Marc Millas wrote: > > you say "extract the data you need" > > That is exactly the point of my question, as the PITR step was obvious. > > How to guess "what is the data" I need ?? > > Well, you asked for the contents of a table AS OF TIMESTAMP <something>. > > That means you know which table and which timestamp. > > So restore the PostgreSQL cluster to that timestamp, connect > and SELECT from the table. > > Yours, > Laurenz Albe > -- > Cybertec | https://www.cybertec-postgresql.com > <https://www.cybertec-postgresql.com/> >
Are you by any chance logging sql at a sufficient level of detail to allow a restore to PIT and sql replay?
