Re: [Gnumed-devel] doc_med , txt blob

[Karsten Hilbert]

On Sun, Sep 17, 2006 at 12:34:59AM +1000, Ian Haywood wrote:

> Ok, you can do
> select * from doc_obj where fk_intended_reviewer = <me> and
>  not found (select 1 from reviewed_doc_obj where fk_doc_obj = doc_obj.id)
Or rather

select * from blobs.v_obj4doc_no_data
where
        pk_intended_reviewer = <me> and
        reviewed is False;

Or perhaps

select * from dem.v_provider_inbox
where
        pk_staff = me and
        type = 'review docs';


> to display an 'Inbox' of unread blobs.

> However after a few years the number of doc_objs the user has ever seen,
> which then have to be run through the sub-query, gets very large.

Doing

select 1 from blobs.reviewed_doc_objs where
        fk_doc_obj = doc_obj.id and
        fk_review != <me>

should be rather fast no matter how many
documents given appropriate indexes and PostgreSQL
cofiguration.

> Having said that, I'm very impressed with the performance of the 0.3 client,
Good. Almost all of it is likely due to improvements in
PostgreSQL performance, wxPython improvement and such
things.

Karsten
-- 
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346