[Xapian-discuss] identifying matched terms

[Olly Betts]

On Tue, Dec 20, 2005 at 04:28:45PM -0800, Jason Toffaletti wrote:
> RuntimeError: InvalidOperationError: positionlist_begin not supported
> 
> Which is consistent with the code I see here:
> http://www.xapian.org/docs/sourcedoc/html/termlist_8h-source.html

That's just the default implementation in the base class.  We implement
positionlist_begin in the subclasses when appropriate (sometimes it
isn't - for example the listed of terms ignored as stopwords is
available from the QueryParser as a TermIterator, but there's no
document context there so positionlist_begin() doesn't make sense
at the conceptual level).

But it sounds like we've missed a case where it should work.  I'll
take a look, but as a workaround you can get a positionlist from the
database by supplying the term and document id explicitly:

http://xapian.org/docs/apidoc/html/classXapian_1_1Database.html#a13

For the backends we currently have, that's as efficient as getting it
from the TermIterator would be.

Cheers,
    Olly