Take the 2-minute tour ×
Geographic Information Systems Stack Exchange is a question and answer site for cartographers, geographers and GIS professionals. It's 100% free, no registration required.

I'm inserting map data from a GML file with srid EPSG:27700 into a PostGIS database table with srid EPSG:4326. I'm using the ST_Transform function as in this snippet:

 ST_Transform(ST_GeomFromGML('<gml:Point srsName="EPSG:27700">
    <gml:coordinates>527347.0,181346.0</gml:coordinates></gml:Point>'), 4326);

After the data was inserted in the database I tried to find if its geometry intersects with an existing table's geometry using the ST_Intersects function. The function returned FALSE in all cases, but I knew that the two geometry layers should intersect.

So, I loaded the layers in QuantumGIS to see what's happening. The PostGIS layers indeed don't intersect, but when I loaded the GML data directly from the GML file using the original srid (EPSG:27700) the map data correctly appeared on top of the pre-existing layer with srid 4326, with which I knew it should intersect. The problem is that the ST_Intersects function (as well as all the others of the same kind) works only for geometries with the same srid, so I need to solve the transformation issue.

Apparently there's either a problem with the ST_Transform function, or I'm missing something that I should know. I use PostGIS 1.5.3. Any help will be greatly appreciated.

share|improve this question
I tried to duplicate your steps and got the expected results. I started with a polygon layer in 4326, and created a new point layer, also in 4326. Then I INSERTed a point into the layer using ST_Transform(ST_GeomFromGML()), using coordinates from a different CRS, but which should fall inside the poly. It worked correctly. So your method seems correct. –  Micha Jul 23 '12 at 16:18
Are you sure the ST_GeomFromGML function has created the geometry as expected? - does the geometry have EPSG:27700 as its srid. –  CHenderson Jul 23 '12 at 22:31
@Micha Sounds strange, I keep on getting it wrong. Thanks for trying it out! –  Orestis Tsinalis Jul 25 '12 at 15:28
@CHenderson I double-checked it and the srid is definitely correct. –  Orestis Tsinalis Jul 25 '12 at 15:28
What I did in the end was to convert the other layer into EPSG:27700 (it's a shape file that I import in PostGIS using the shp2pgsql tool), and to import the GML without transforming the geometry to a different spatial reference system. It then worked as expected, but I would really like to understand what went wrong in the first case. –  Orestis Tsinalis Jul 25 '12 at 15:28
show 3 more comments

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.