[CWB] CWB: problems with indexing a corpus

[Stephanie Evert]

> I am trying to process a vrt file with cwb-encode.
> The file has pos tagging and I used the examples from CWB-manuals as a template. 
> 
> I run the command
> cwb-encode -f /path_to_file.vrt -d /path/datafiles -R /path/registry/corpus_name -9 -c utf8 -P pos -P lemma -S text -S p -S s
> 
> and I am getting the following warnings:
> > Annotations of s-attribute <text> not stored (file /xxx.vrt, line #1, warning issued only once).
> > Annotations of s-attribute <p> not stored (file /xx.vrt, line #3, warning issued only once).
> > Annotations of s-attribute <s> not stored (file /xx.vrt, line #4, warning issued only once).

That means the start tags of these XML elements contain attribute-value pairs, which you're ignoring – cwb-encode simply warns you about this fact. 

> And the programme terminates without producing any result. 

That sounds like an error, though, and completely unrelated to the warnings.  After successful completion of the command, your data directory /path/datafiles should be populated with index files.

Are you sure there isn't any error message?

A first step would be to re-run cwb-encode with the -v option added (at the start, not after the attribute flags). This should print how many tokens have been read and encoded from the vrt file.

Best,
Stephanie