Thread: FTS on patterns
Hello,
In our application we need a pattern based text search, that is if user supplies 'foo bar' he receives only documents containing words that begin with 'foo' or 'bar', including themselves (something similar to MySQL '+foo* +bar*' text search query). Currently we use separate table for unique words in document and do LIKE 'foo%' search on that table, combining the results afterwards, and I am not really happy with that solution because of its speed and overcomplication of the queries involved. Does anyone know of a better approach? Could this be achieved with Postgres full text search?
In our application we need a pattern based text search, that is if user supplies 'foo bar' he receives only documents containing words that begin with 'foo' or 'bar', including themselves (something similar to MySQL '+foo* +bar*' text search query). Currently we use separate table for unique words in document and do LIKE 'foo%' search on that table, combining the results afterwards, and I am not really happy with that solution because of its speed and overcomplication of the queries involved. Does anyone know of a better approach? Could this be achieved with Postgres full text search?
tsearch can help you; i have a system with it ... and its ok. El mar, 05-08-2008 a las 18:03 +0400, Vyacheslav Kalinin escribió: > Hello, > > In our application we need a pattern based text search, that is if > user supplies 'foo bar' he receives only documents containing words > that begin with 'foo' or 'bar', including themselves (something > similar to MySQL '+foo* +bar*' text search query). Currently we use > separate table for unique words in document and do LIKE 'foo%' search > on that table, combining the results afterwards, and I am not really > happy with that solution because of its speed and overcomplication of > the queries involved. Does anyone know of a better approach? Could > this be achieved with Postgres full text search? > > >
"Vyacheslav Kalinin" <vka@mgcp.com> writes:
> In our application we need a pattern based text search, that is if user
> supplies 'foo bar' he receives only documents containing words that begin
> with 'foo' or 'bar'
There's support for prefix text search in CVS HEAD, but not in any
released version ...
regards, tom lane
On Tue, Aug 5, 2008 at 6:53 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
There's support for prefix text search in CVS HEAD, but not in any
released version ...
regards, tom lane
Oh, does that mean it won't be released until 8.4?
Vyacheslav Kalinin wrote: > On Tue, Aug 5, 2008 at 6:53 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote: > >> There's support for prefix text search in CVS HEAD, but not in any >> released version ... >> >> regards, tom lane > > Oh, does that mean it won't be released until 8.4? New features are never introduced in minor releases (8.3.x) only bugfixes. -- Richard Huxton Archonet Ltd