Hi Simon,
On 20/2/2012 12:26 AM, Simon Slavin wrote:
You may get only marginally slower results by having just an index on
text_field. It depends on how 'chunky' your values are: whether you have lots
of rows with the same value in [text_field] but different values in [category],
for example. Try some timing comparisons of the current setup with one after
deleting all those indexes and creating just one for text_field.
But the great advantage of using FTS is in searching text fields for contents.
If you're searching for exact contents of a text field FTS is a waste of time
and filespace.
That is a very good point! In our case, we have a very large spread of
data... so, each text_field appears in less than 5 rows! So, you're
right - if we do index just text_field, there would be a scan of just 5
rows or so to find the correct answer!
Thanks, we'll try this.
Best Regards,
Mohit.
20/2/2012 | 12:58 AM.
_______________________________________________
sqlite-users mailing list
[email protected]
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
