hi mathieu, i cannot split the data in a very meaningful way. My example is a bit misleading. basically, I can have the field FILTER indexed with 12K or 13B-14K and the user might want results that have field FILTER > 13 and K, the first one (12K) won't match because 12 < 13 but the second will (it has in the string 14K which is 14 > 12 and K). So I don't think that I can alter the way I indexed data to be able to use facet .. it is just not possible .. I am just trying to have people use regular free text search but then filter the results using a regex ..
thanks. On 4/12/08, Mathieu Lecarme <[EMAIL PROTECTED]> wrote: > > hi, > > > > I have a question ... I need to be able to filter a search using a > > regex. I > > cannot used facet as the filtering is pretty complex (but easy to > > perform > > using a regex). > > For instance I have stored in the field ID the value 12G and I want to > > basically filter out all the results that are > 12 with G so for > > instance > > 14G will match but 8G and 14B would not. Using a regex this is simply > > "[1-9]+[3-9]G" .. > > i am wondering what the right approach is to tackle such a situation .. > > > > thanks. > > > regex match is only useful when you first select a prefix, wich is a basic > lucene feature : put the pointer just up to the first term begining with > "toto". > Your query don't have any prefix. > What happen if you split your data in two field "12" and "G", "14" and > "B", or, better, if it's number, "12G" can be indexed as "12000000"? > > M. >
