Unfortunately, the keys can not be predefined or fixed. it is a doc, the reason jsonb
is used. It works well for small docs with small number of keys.
but really slow with large number of keys. If this issue is resolved, I think Postgresql
would be an absolutely superior choice over MongoDB.for document data.
On Sun, Nov 29, 2015 at 12:37 AM, John R Pierce <pierce@hogranch.com> wrote:
On 11/28/2015 6:27 PM, Tom Smith wrote:
Is there a plan for 9.6 to resolve the issue of very slow query/retrieval of jsonb fields when there are large number (maybe several thousands) of top level keys. Currently, if I save a large json document with top level keys of thousands and query/retrieve field values, the whole document has to be first decompressed and load to memory before searching for the specific field key/value.
If it was my data, I'd be decomposing that large JSON thing into multiple SQL records, and storing as much stuff as possible in named SQL fields, using JSON in the database only for things that are too ambiguous for SQL.