> 1. Currently we use Verity and have more than 20 collections, each collection > has a index for public items and a index for private items. So there are > virtual collections which point to each collection and a virtual collection > which points to all. For example, we have AA and BB collections. > > AA virtual collection --> (AA index for public items and AA index for private > items). > BB virtual collection --> (BB index for public items and BB index for private > items). > All virtual collection --> (AA index for public items and AA index for > private items, BB index for public items and BB index for private items). > > Would you please tell me what I should do for this if I use Solr?
There are multiple ways to solve this, depending on the nature of your collections. If they have somewhat different schemas, a natural choice would be to make multiple cores: AA-private, AA-public, BB-private, BB-public. Now you can query them individually or in combinations through the shards parameter. From next Solr version you can use virtual collections for the shard parameter, e.g. &shards=AA,BB etc. (See http://wiki.apache.org/solr/SolrCloud#Distributed_Requests) If all your content is (roughly) the same kind of data, you could also solve your virtual collection issue through a "collection" field in your schema, and simply select collection through filters: &fq=collection:AA. You could even write a Search Component which translates a &collection= parameter in the request into the correct filters if you want to hide this implementation to the front ends. > 2. Our project has different kind format files I need index them. For > example, xml files, pdf files and text files. Is it possible for Solr to > return a search result from all? Sure. PDF and text files can be indexed through the ExtractingRequestHandler. XML can be indexed from XMLUpdateHandler or DataImportHandler. Solr uses Apache Tika internally to extract text from PDFs and other rich document formats. > > 3. I got a error when I index pdf files which are version 1.5 or 1.6. Would > you please tell me if there is a patch to fix it? How did you try to index these PDFs? What version of Solr are you using? Exactly what error message did you get? -- Jan Høydahl, search solution architect Cominvent AS - www.cominvent.com Training in Europe - www.solrtraining.com
