On Thu, Jul 9, 2009 at 8:59 AM, Düster Horst<[email protected]> wrote: > Martin > I found out further issues related with OID Indexes (1) and table versioning > (2) and one enhancement (3). > 1. > When I select a table with OID index, the meta data Indexes part shows: > Name Column(s) > afu_prog05_oid_idx ??? (#-2) > afu_prog05_idx wkb_geometry
Ah, tables with oids, that's something I've completely forgotten... Anyway, isn't it discouraged to use oids for regular tables? > 2. > * The schema select box content should be sortet. > * The geometry index and all other indexes of the versionized table should > be a patial index with the constaint "where time_end = NULL ". With a > partial index you will have huge performance advantages. With the time the > amount of archived or historized records will grow up. P.e. one table of our > cadastral data contains more than 1 Mio records, but 700000 of them are not > in production state (end_time <> NULL). With the partial index these records > don't exists for Postgres, as the result you do not have any performance > problems with your production data. That's why I'm not adding any indexes to the tables with versioning, because I wasn't sure how to do it correctly. I was thinking about using multicolumn indexes - e.g. geometry+endtime, that would probably do the right thing too. But the partial indexes seem to be more appropriate. Thanks for sharing. > 3. I would be nice to add a selected table to Qgis mapcanvas. Already on my TODO :-) Bye Martin _______________________________________________ Qgis-user mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/qgis-user
