Thanks Marbex, Puneed and specially to Scott for the snippet() tip.
'ined' is Porter stemmed to 'in'! :-)
So that is why it finds so many rows.
Now, to find 17081 rows from a table with 25672, fts3 took 4.47 seconds!
The query is:
select title from fts_news where description match 'in'
If
Right, Puneet, this is what I mean.
I tried the same query with another database with 25.600 rows:
select * from fts_news where description match 'ined'
the query returns 17.080 rows which is wrong because If I do:
select * from fts_news where description like '% ined %'
The query retu
On Sat, Sep 26, 2009 at 5:31 PM, P Kishor wrote:
> On Sat, Sep 26, 2009 at 6:36 PM, marbex wrote:
>> Shopsland gmail wrote:
>>> select title from fts_news where fts_news match 'ined'
>>
>> It looks that you only want to query the title field, so the query should
>> be:
>> select title from fts
On Sat, Sep 26, 2009 at 6:36 PM, marbex wrote:
>
>
> Shopsland gmail wrote:
>>
>> select title from fts_news where fts_news match 'ined'
>>
>
> It looks that you only want to query the title field, so the query should
> be:
> select title from fts_news where title match 'ined'
> or
> select tit
Shopsland gmail wrote:
>
>select title from fts_news where fts_news match 'ined'
>
It looks that you only want to query the title field, so the query should
be:
select title from fts_news where title match 'ined'
or
select title from fts_news where fts_news match 'title:ined'
--
View this
Hi,
Mining my fts3 table stemmed with Porter I found that there is a 'phantom'
token indexed by fts3.
When I issue this query:
select title from fts_news where fts_news match 'ined'
I got about 700 rows that do not have the word 'ined' in any form (ie: ined
or inedible).
If I issue those que
6 matches
Mail list logo
