Thanks for your reply Why I care the language: according to the documentation: "A term is a contiguous sequence of eligible characters, where eligible characters are all alphanumeric characters and all characters with Unicode codepoint values greater than or equal to 128. All other characters are discarded when splitting a document into terms. Their only contribution is to separate adjacent terms." if the path include folder or file name in France or Hebrew or Arabic and the user what to search that word according to above the FTS will not found it. Am I wrong?
about the backslash/slash if I have a path like this and the user search for עברית did the FTS will find it? c:\folders\video\עברית about the creation I was confuse it not temporary its virtual so ignore my question Thanks On Sat, Jan 12, 2013 at 5:58 PM, Michael Black <mdblac...@yahoo.com> wrote: > I'm not sure I understand your problem. > Why do you care what language it is? Aren't you just wanting to tokenize > on > backslash? > > Simple way is to replace all spaces in the path with another char (e.g. > '_') > then replace all backslashes with a space. > Then you can just use the default tokenizer and make the same changes on > any > user queries. So you map the user query to your storage format. > > Not sure why you would want to use some stem tokenizer on paths. > > And, of course, you're reinventing the wheel unless you have some special > purpose in mind. > > http://locate32.cogit.net/ > > -----Original Message----- > From: sqlite-users-boun...@sqlite.org > [mailto:sqlite-users-boun...@sqlite.org] On Behalf Of moti lahiani > Sent: Saturday, January 12, 2013 4:37 AM > To: sqlite-users@sqlite.org > Cc: Moti LAHIANI > Subject: [sqlite] FTS questions > > Hello all > > I'm new with sqlite3 and sql. > > I have data base that include path columns (file system path like c:\bla > bla\myFiles\1.txt) > > On that columns I need to do > > 1) search for patterns in case the user want to find a file or > directory > > 2) search for prefix path in case the user rename a file or directory > > the Database include about 1.5-2.5 million records and to use the "LIKE" is > not possible because the result time. > > As an alternative I want to use FTS3 or FTS4 but I think I have a problems > with what I read here: http://www.sqlite.org/fts3.html#section_1_4 > > And here: http://www.sqlite.org/fts3.html#section_6_3 > > I need to specify the language to FTS to use it as tokenize but the path > can include multi languages how can I configure the FTS table to use all > languages > > How can I tell to FTS to token the path only according to the character "\" > ? > > > > More than that when creating FTS table it creates with the TEMPORARY key > word. My question is: do I need to create this table each time I run the > data base (because the temporary word) or for each connections (in case > of multiconnections) or this is a table like all tables I declared and it > stay in the data base even if I restart my PC > > > > Thanks a lot > _______________________________________________ > 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 > _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
