bram wrote:
> I tried to make a connection to a postgresql-table within pmapper1.0.8 
> as well al in pmapper2 (both dated 2006-02-20). I work on a 
> windows-machine with Firefox. The connections succeeded, the masp were 
> drawn, but when retrieving  info with  'identify' or with 'select'  we 
> see the correct number of lines in the resulttable and the correct 
> headers but without any data and the polygon cannot be selected. The 

I would guess that you do not specify the correct field names: they
should be in *lower* case, in the database and in the 'RESULT_FIELDS'
metadata. Try to comment out 'RESULT_FIELDS', then you should get *all*
fields.
A full description of the layer would help, too.

> postgreql-tables instead are containing values. After that the map tries 
> to be redrawn but it sticks at:   *Fatal error*: Object expected as 
> argument. in *C:\ms4w\apps\pmapper-1.0.8\incphp\map.php* on line *568 
> *(in 1.0.8), but in 2 only the the 'loading-sign' appears without any 
> success.
> I was wondering if Pmapper was yet postgis-proof of if I was doing 
> something wrong. Who can give me a kind of solution?
> 
> This is the in the map-file:
> CONNECTIONTYPE POSTGIS
> CONNECTION "dbname=project1 host=localhost user=postgres password=abcd 
> port=5432"
> DATA "wkb_geometry from water1 as foo1 using unique wkb_geometry using 
> SRID=-1"
> 

I doubt that it is working with the geometry field as unique field. The
result takes the index for that row to create a polygon for the dynamic
resultlayer, and I guess it uses by default the field defined as UNIQUE.
  In this case also the '\d water1' output from psql could help. I
typically use the OID field, and currently attribute searches in pm work
only if the OID field exists (that was always the case until somebody
decided to drop this default field in PG 8.1).

I am using PostGIS layers quite a lot because of their
flexibility. So yes, I would call p.mapper 'postgis-proof'.


armin

Reply via email to