Oracle Text - wildcard expansion too many rows [message #312660] |
Wed, 09 April 2008 10:00 |
SaraC
Messages: 81 Registered: August 2005
|
Member |
|
|
Just typed a big message and lost it all so will try to keep this shorter.
Using Oracle Text 10g, implementing a CONTAINS search. Have created a stoplist using ctx.ddl however it doesnt get any terms added to it since we want to search on all terms including stopwords. so all terms are indexed for searching. ctxsys.context Index has been created, with printjoins specified so special chars are indexed too.
I want to search for all terms in any order.
Trying to run a search using
CONTAINS(term, '(%pain% AND %in% AND %ear%)
gets the message Wildcard query expansion returns too many results error - due to the use of 'in' (ive tried the other terms without in it returns results.
'In' matches a very large number of terms in our dataset >15000 ( there is no max rows defined explicitly as I thought the default was 5000)
The wierd thing is that when I use
CONTAINS(term, '(%pain% AND %the% AND %chest%)
where %the% on its own matches well over 15000 records, a result is returned with no error.
Is this a problem with %in%?
Thanks for any info
Sara
|
|
|