[gdal-dev] Anyone against making OGR SQL 'SUBSTR' conformant with other SQL implementations ?

Even Rouault even.rouault at mines-paris.org
Thu Nov 24 16:47:04 EST 2011


The implementation of the OGR SQL SUBSTR(string, offset[, length]) function 
considers that the character offset is like the convention of the C language, 
that is to say that the first character is offset 0, the second one is offset 1, 
etc... All other SQL implementations I've found consider 1 as the first 
character, 2 as the second one.

For details and patch, see http://trac.osgeo.org/gdal/ticket/4348

The only reason I'm reporting publicly here is that fixing that could break 
existing applications/scripts. The risk seems limited however because this 
function was introduced in 1.8.0 and happens not to be documented in 

So I think I'll apply the fix in trunk, and likely in 1.8 branch too. With a 
notice in the backward compatibility section of the NEWS file.


More information about the gdal-dev mailing list