that has answered my concern about the index size/duplicated data. but the other one is about presenting the search results, results should be one with list of files. so in this case I would need to write some logic before showing the results right? (may be like comparing each result solrdocument/xml element with some field value if that matches with the previously populated document, then avoid the duplicated result but just add the file name/path to the previous entry.
I wanted to avoid this, that's where I was checking if there is any hidden solution for this in SOLR. May be a kind of wrapper would help all others for these kind of requirements. Best Regards, Kranti K K Parisa On Tue, Feb 16, 2010 at 9:02 PM, Erick Erickson <erickerick...@gmail.com>wrote: > Unless you have *evidence* that the indexing each pdf with > the form data as a single SOLR document is a problem, > I would just index the fields with each document rather > than try to index the PDFs as multivalued. The space > used by duplicating the form field data is probably a > tiny fraction of the data you'll index and the space savings > are probably not worth the complications. > > Erick > > 2010/2/16 Kranti™ K K Parisa <kranti.par...@gmail.com> > > > Hi, > > > > When we index using SOLR, we have an option called multivalued. How does > > that work with multiple files associated with same document. > > > > For example: submiting a form with some fields + list of pdf files > > index process: > > 1) considering all the form fields as individual solr input document > fields > > (using setField) > > 2) getting contents of the pdf files and addiing them to a field called > > contents which is declared as multivalued (using addField) > > > > So now when we search for something, and assume the search text has been > > found in the second file, how can we display the file name? > > I know we have to store the file name also as a field to get the file > > names, > > but how does that work with multivalued fields. > > > > Is there any way to achieve this? One way could be indexing solr input > > documents for each file, but my concern is that would duplicate the other > > form fields data. > > > > Best Regards, > > Kranti K K Parisa > > >
