[pgrouting-users] Performance problems with pgr_dijkstra

[Daniel Kastl]

>
>
> I don't understand, why the procedure lasts so much longer at the map of
> Europe although both
> temporary routing tables are almost identical in size and don't depend on
> the rest of the databases
> (no joins or other complicated stuff). Maybe someone has an idea and can
> point me in the right
> direction.
>
>
Hi Eric,

First of all I'm not sure, if I really understand what you mean with
"creating a temporary table". Did you run "CREATE TEMP TABLE ..."?

You're right, that with pgRouting the amount of data selected from the
network table matters. And the fastest way to select only a part of the
network table is by selecting a bounding box. You should have an index on
your geometry column as well. Then you don't need to create temporary
tables.

Back to your question: as far as I remember, the size of ID's can matter. I
experienced this when I used data, that had already source and target ID's
in place, which all had the same number of digits. Renumbering (starting
from 1) helped to improve the speed. Though I can't tell this is the reason
in your case.

Have you also run VACUUM and ANALYZE on your table?

Regards,
Daniel



-- 
Georepublic UG & Georepublic Japan
eMail: daniel.kastl at georepublic.de
Web: http://georepublic.info
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/pgrouting-users/attachments/20150211/449cebc5/attachment.html>