performance - Comparison of JSON and User-Defined types in Postgres 9.3 -
i wonder why there fuss json support in postgres 9.3. advantages of json on user-defined types (udts)? pitfalls in using udts? access tables udts inefficient? alter type add attribute slow? how udts physically stored postgres?
please, explain , give links additional information.
as roman pekar mentioned in 1 of previous answers, json support offer more flexibility , offers possibility kinda mimic nosql behavior on relational one.
furthermore, makes easier in client-server applications store json values sent client directly database.
one can use 30% of fields client of application, 30% another, , on, not having define multiple tables or tables large set of columns. thus, 1 can store large chunks of heterogeneous information 1 place.
last not least, json standard , it's supported many of big programming languages.
(we using feature in our project (and have been using since in beta); furthermore, main reason chose postgres our application, needed big db decoupled information. tried using nosql databases needed many tables store information in, , costly on "joins". on other hand, have been hard cope relational db, instead of going half-relational half-nonrelational, chose postgres's json support.)
Comments
Post a Comment