[postgis-users] Numeric overflow when inserting large polygons

Hartmut Tschauner tschaun at fas.harvard.edu
Mon Dec 22 10:37:05 PST 2003

Thanks so much for the hint, which was dead on. I had considered this
possibility for a moment, but had quickly discarded it as too much of a
coincidence. In retrospect, the error message was perfectly clear and there
wasn't any coincidence at all: the offending fields were area and
circumference measurements of the shapes and large polygons tend to have
large areas and circumferences. The loader apparently does not scan the
values to make sure the columns of the new table are properly dimensioned
for all the data items.

I'm glad this problem had an easy fix. I like PostGIS quite a bit, but
difficulties with large geometries would have been a serious problem for my

Thanks again for your help, it is greatly appreciated.

Happy holidays,

Hartmut Tschauner
Dept. of Anthropology
Harvard University
Peabody Museum
11 Divinity Ave.
Cambridge, MA 02138
ph. (617) 384-9803
fax (617) 496-8041

-----Original Message-----
From: strk [mailto:strk at keybit.net] 
Sent: Monday, December 22, 2003 9:26 AM
To: PostGIS Users Discussion
Cc: tschaun at fas.harvard.edu
Subject: Re: [postgis-users] Numeric overflow when inserting large polygons

tschaun wrote:
> Dear listers,
> I have just finished installing PostgreSQL 7.4 and PostGIS 0.8 on a RedHat
> machine. The installation seems to have gone fine and everything usually
> works as expected. However, I'm having trouble with a few very large
> polygons in my test layers that cause a numeric overflow error when I try
> enter them into PostGIS tables using SQL code generated by shp2pgsql.
> "ERROR: numeric field overflow. DETAIL: The absolute value is greater than
> or equal to 10^8 for field with precision 19, scale 11."

I think this has nothing to do with the geometry field, but rather with
one of the associated values (attributes in the .dbf file).
Open the output of shp2pgsql and see how numeric fields are defined in
the CREATE TABLE statement.


More information about the postgis-users mailing list