Hello Vitaly > -----Original Message----- > From: Vitaly Burovoy [mailto:vitaly.burovoy@xxxxxxxxx] > Sent: Montag, 25. Januar 2016 14:25 > To: Christophe Pettus <xof@xxxxxxxxxxxx>; clavadetscher@xxxxxxxxxxxx > Cc: Postgres General <pgsql-general@xxxxxxxxxxxxxx> > Subject: Re: ERROR: check constraint - PostgreSQL 9.2 > > On 1/24/16, Christophe Pettus <xof@xxxxxxxxxxxx> wrote: > > > > On Jan 24, 2016, at 9:01 PM, Charles Clavadetscher > > <clavadetscher@xxxxxxxxxxxx> wrote: > > > >> What is the point of having a check constraint that is not checked? > > > > Well, it *is* checked going into the future; it's just not checked at the > > time the constraint is added. Ultimately, you do want to fix the data, but > > this makes it a two-step process, and reduces the time the table is locked > > against access. > > NOT VALID constraint checks new and updated rows, and gives an extra > time to fix current data and be sure there will be no new rows that > violates the check constraint during and after the fixing process. > > N.B.: Prior 9.4 it does *NOT* reduce the time the table is locked > because VALIDATE CONSTRAINT requires ACCESS EXCLUSIVE[1] and uses > seqscan for check table's rows. > > P.S. Lucas, If you have not received answers, you can find all of them > as a thread by the link: > http://www.postgresql.org/message-id/flat/CAE_gQfXTns1FR5Fx9wxpo1oZYwat639ua-gAqWZyNg201HCU=Q@xxxxxxxxxxxxxx > > P.P.S.: Christophe, Charles! Please, use "Relpy to all" to be sure > the sender gets your answers even if he haven't subscribed to the > mailing list. Oops. Honestly I did not think of that. I will keep that in mind in the future. Thank you for the hint. Charles > > [1]http://www.postgresql.org/docs/9.2/static/sql-altertable.html > -- > Best regards, > Vitaly Burovoy -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
