Re: [GENERAL] Table create time
От | Achilleas Mantzios |
---|---|
Тема | Re: [GENERAL] Table create time |
Дата | |
Msg-id | fd5aa2a7-2ea3-3354-84ec-80e09e07c54a@matrix.gatewaynet.com обсуждение исходный текст |
Ответ на | Re: [GENERAL] Table create time (Achilleas Mantzios <achill@matrix.gatewaynet.com>) |
Список | pgsql-general |
On 31/08/2017 16:12, Achilleas Mantzios wrote: > On 31/08/2017 14:03, hamann.w@t-online.de wrote: >>>> On 31/08/2017 09:56, hamann.w@t-online.de wrote: >>>>> Hi, >>>>> >>>>> is there a way to add a table create (and perhaps schema modify) timestamp to the system? >>>>> I do occasionally create semi-temporary tables (meant to live until a problem is solved, i.e. longer >>>>> than a session) with conveniently short names. >>>> In FreeBSD you'd do smth like this to find the file creation time : >>>> ls -lU <path to your cluster>/data/PG_9.3_201306121/16425/12344 >>>> >>>> where 12344 is the filenode of the relation in question. In ext4 you may do this albeit with more difficulty. >>>> >> Hello Achilleas, >> >> many thanks for responding. There are two problems; >> a) accessing the filesystem will likely require some extra effort (e.g. installing an untrusted programming >> language) > No need for this. You may use builtin pg_stat_file function . I see it supports a "OUT creation timestamp with time zone"parameter. Sorry, just tested that against both FreeBSD pgsql9.3 and Ubuntu/ext4 10beta3, and .creation returns null in all tests. Soyes you might need to write your own function . >> b) a dump/restore will modify the dates > That would be a problem, but this is not a common use case. Anyways you can always write an event trigger and store somemessage in a log file. This should survive dump/restores . > >> >> best regards >> Wolfgang Hamann >> >> >> > -- Achilleas Mantzios IT DEV Lead IT DEPT Dynacom Tankers Mgmt
В списке pgsql-general по дате отправления: