i would not cross-reference solr results with your database to merge unless you want to spank your database. nor would i load solr with all your data. what i have found is that the search results page is generally a small subset of data relating to the fuller document/result. therefore i store only the data required to present the search results wholly from solr. the user can choose to click into a specific result which then uses just the database to present it.
use data import handler - define an xml config to import as many entities into your document as you need and map columns to fields in schema.xml. use the Wiki page on DIH - it's all there, as well as example config in the solr distro. allistair On Oct 14, 2010, at 6:13 PM, Juan Manuel Alvarez wrote: > Hello everyone! I am new to Solr and Lucene and I would like to ask > you a couple of questions. > > I am working on an existing system that has the data saved in a > Postgre DB and now I am trying to integrate Solr to use full-text > search and faceted search, but I am having a couple of doubts about > it. > > 1) I see two ways of storing the data and make the search: > - Duplicate all the DB data in Solr, so complete results are returned > from a search query, or... > - Put in Solr just the data that I need to search and, after finding > the elements with a Solr query, use the result to make a more specific > query to the DB. > > Which is the way this is normally done? > > 2) How do I synchronize Solr and Postgre? Do I have to use the > DataImportHandler or when I do the INSERT command into Postgre, I have > to execute a command into Solr? > > Thanks for your time! > > Cheers! > Juan M.