[gdal-dev] CopyLayer oddity using GML driver

[Ari Jolma]

On 11/21/2011 08:57 PM, Even Rouault wrote:
> Le lundi 21 novembre 2011 11:20:33, Ari Jolma a écrit :
>> Hi,
>>
>> I'm making a new GML layer from an existing layer. For some reason the
>> CopyLayer method leaves adjusted fields out but CreateField,
>> CreateFeature sequence does not. I guess they should be identical?
>>
>> i.e.
>>
>> my $datasource = Geo::OGR::Open('Pg:...');
>> my $l = $datasource->ExecuteSQL('...');
>> my $gml = Geo::OGR::Driver('GML')->Create('/vsistdout/');
>> $gml->CopyLayer($l, 'a');
>>
>> leaves fields like "a b" out, while
> I guess it is because the GML driver transforms "a b" into "a_b" to make it a
> valid XML name. But the implementation of CopyLayer() uses SetFrom(), which
> cannot guess that "a b" values from source feature should be mapped to "a_b"
> of target feature.
>
>> my $datasource = Geo::OGR::Open('Pg:...');
>> my $l = $datasource->ExecuteSQL('...');
>> my $gml = Geo::OGR::Driver('GML')->Create('/vsistdout/');
>> my $l2 = $gml->CreateLayer('test');
>>
>> my $d = $l->GetLayerDefn;
>> for (0..$d->GetFieldCount-1) {
>>       my $f = $d->GetFieldDefn($_);
>>       $l2->CreateField($f);
>> }
>>
>> $l->ResetReading;
>> while (my $f = $l->GetNextFeature) {
>>       $l2->CreateFeature($f);
>> }
>>
>> works ok (and prints out warnings about the field name adjustments)
> Works ok perhaps, but it is potentially risky. You should not use $l2-CreateFeature($f) with $f being a feature that has the feature definition of $l and not $l2. That could crash in some situations... You're lucky here that it works for you because you have the same number of fields and same type.

Hm, after some thinking I think the best approach is to construct the 
original SQL to assign XML-valid column names. The same change 
can/should be done in for the WFS DescribeFeature response. The the 
CopyLayer method should work ok (seems so after initial tests).

Cheers,

Ari