Re: lifetime of the old CTID
Adrian Klaver <adrian.klaver@aklaver.com>
From: Adrian Klaver <adrian.klaver@aklaver.com>
To: Matthias Apitz <guru@unixarea.de>, Ron <ronljohnsonjr@gmail.com>
Cc: pgsql-general@lists.postgresql.org
Date: 2022-07-06T20:15:09Z
Lists: pgsql-general
On 7/6/22 12:51, Matthias Apitz wrote: > El día miércoles, julio 06, 2022 a las 02:33:42p. m. -0500, Ron escribió: > >> On 7/6/22 01:18, Matthias Apitz wrote: >> [snip] >>> Ofc, each table has its own primary key(s), used for example for the >>> SELECT ctid, * FROM d01buch WHERE ... >>> >>> As I said, we came to PostgreSQL from Sybase (and Oracle) and Sybase has >>> for each table a so called SYB_IDENTITY_COLUMN which is static for the >>> table and its value does not change. >> >> The DBA who designed that should be flogged for pretending that >> SYB_IDENTITY_COLUMN is a primary key. > > Now things are coing to blaming. Nobody said that SYB_IDENTITY_COLUMN > is a primary key, but it is uniqu to identify a row in a table once > known. Which is the definition of a PRIMARY KEY. At any rate as has been noted repeatedly ctid is not that. > > matthias -- Adrian Klaver adrian.klaver@aklaver.com