Re: [HACKERS] MAIN, Uncompressed?
От | Simon Riggs |
---|---|
Тема | Re: [HACKERS] MAIN, Uncompressed? |
Дата | |
Msg-id | CANP8+jJoOJDdR+ej2534FxueO=V2cq91KMPPiWYXD7ArkiXj4w@mail.gmail.com обсуждение исходный текст |
Ответ на | Re: [HACKERS] MAIN, Uncompressed? (Tom Lane <tgl@sss.pgh.pa.us>) |
Ответы |
Re: [HACKERS] MAIN, Uncompressed?
|
Список | pgsql-hackers |
On 25 August 2017 at 14:08, Tom Lane <tgl@sss.pgh.pa.us> wrote: > Simon Riggs <simon@2ndquadrant.com> writes: >> On 25 August 2017 at 13:21, Tom Lane <tgl@sss.pgh.pa.us> wrote: >>> If you know compression isn't useful, but you don't want to fail on >>> wide values, then "external" should serve the purpose. > >> Well, almost. External toasts at 2048-ish bytes whereas Main toasts at >> 8160 bytes. >> The rows are typically near 4kB long, so if marked External they would >> always be toasted. >> It's desirable to have the full row in the heap block, rather than >> have to access heap-toastindex-toastblocks in all cases. >> The data is also incompressible, so Main just wastes time on insert. >> Hence, we have a missing option. > > Maybe, but the use case seems mighty narrow. JSON blobs between 2kB and 8160 bytes are very common. String length is maybe a poisson distribution, definitely not uniform. -- Simon Riggs http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
В списке pgsql-hackers по дате отправления: