[postgis-users] Collect vs. ST_Collect GeomUnion vs ST_GeomUnion

Mark Cave-Ayland mark.cave-ayland at ilande.co.uk
Wed Aug 22 05:18:24 PDT 2007


On Wed, 2007-08-22 at 07:58 -0400, Obe, Regina wrote:
> Hi Mark,
> 
> I used your install exe.  The recreated template_postgis() is right and
> has those aggregate functions.  At first I thought maybe I ran the wrong
> lwpostgis_upgrade.sql file after upgrading even though the version
> changed from "needs upgrade". So I verified the upgrade file had changed
> and reran again.
> 
> Looking at the upgrade file - I don't see definition of the new
> aggregates, although as you said those aggregates do appear in
> lwpostgis.sql file  
> 
> 
> Here are the details of my install
> 
> Before the installation I was running 1.2.1
> 
> 
> 1) Ran your install. (well it failed first time because presumably I had
> something running using the dlls)
> 2) Restarted postgresql service
> 3) Reran install - ran okay this time - said it would drop
> template_postgis responded yes
> 4) Completed successfully
> 5) Restarted my postgresql service
> 6) Did a select postgis_full_version() on the db I had - I forget what
> exactly it said but it did say "need upgrading"
> 6) Ran the lwpostgis_upgrade.sql file in the
> share/contrib/postgis/lwpostgis_upgrade.sql ( the file is dated
> 8/18/2007 5:13 AM which I assume is when you built it).
> 
> Now the first time I ran this I was actually going thru terminal
> services so I thought maybe I had PgAdmin open on my local pc instead of
> the server I was upgrading, but it did get rid of the need upgrading
> part after install so I presumed it was okay and changed to
> "POSTGIS="1.3.1" GEOS="3.0.0rc4-CAPI-1.3.3" PROJ="Rel. 4.5.0, 22 Oct
> 2006" USE_STATS"
> 
> 7) Ran the tests I mentioned which failed
> 8) Reran the upgrade file verifying I had the correct one.
> 
> 
> Thanks,
> Regina


Hi Regina,

If you look at the latest SVN
http://svn.refractions.net/postgis/trunk/utils/postgis_proc_upgrade.pl,
I think that the commented out code section explains the problem. So
looks as if this is now going to become more of an issue now that the
aggregate names have changed :(

I think the answer would be to alter the postgis_proc_upgrade.pl so that
it searches the PostgreSQL catalogs first, and only drop the aggregates
that already exist in the database. I'll add it to the TODO list, but no
guarantees as to when it will get done - although you'll be the first
person on my list to ask when it needs testing ;)


ATB,

Mark.

-- 
ILande - Open Source Consultancy
http://www.ilande.co.uk





More information about the postgis-users mailing list