There is not currently any support for this in fts. Put another way, you can count the words like you'd count them without fts :-).
Long-term, I do plan to add support for figuring out some meta-information about tokens across the index, but there's nothing really driving that at this time. -scott On Tue, Jun 10, 2008 at 7:24 PM, Tito Ciuro <[EMAIL PROTECTED]> wrote: > Hello, > > I was reading the document 'http://www.sqlite.org/cvstrac/wiki?p=FtsUsage' > and I was wondering whether there is a way to count the number of > times a word is present in text indexed with FTS 3. > > > For example: > CREATE VIRTUAL TABLE recipe USING fts3(name, instructions); > INSERT INTO recipe VALUES ( 'coffee cake', 'Bake in oven for 30 > minutes. Remove from the oven and allow to cool for 5 minutes. Turn > pan over onto a serving platter to remove. Serve by pulling apart > chunks with forks' ); > > Is there some way to count the times the word 'oven' appears in the > text? Where I'm trying to get at: to retrieve all rows where 'oven' is > more often referenced. This would allow me to build a"relevance" list, > for example. > Thanks in advance, > > -- Tito > _______________________________________________ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users > _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
