How large are your individual tables, and how many possible joins are there?
The general advice for schemas (Solr and Lucene) is to use one per database and denormalize denormalize denormalize. Solr&Lucene are very fast even with hundreds of millions of (small) documents. The new Group Collapsing feature makes it efficient to completely denormalize your database and then "collapse" result sets back to the efficiencies of multiple DB tables. Having different languages in different fields works quite well. With the <copyField> directive in schema.xml it is easy to parse them for different languages and then make a common language-agnostic field. On Wed, Sep 28, 2011 at 5:43 AM, Cynthia Dematteis-Krug - Xsite GmbH < [email protected]> wrote: > Good morning, > I am new in Solr and i do not find any technical support to ask, only this > address. I hope i am ok here. > I used to work in Lucene and we created there one index per table. This > way, if we wanted to search products, we use the "Products" index, if we > wanted to search in customers, we used the "customer" index and so on. We > had different indexes per language as well, like: customer_de, product_en, > ect.. > Now we would like to migrate everything to Solr which i think is a good > idea, but i realized that every solar Server works with one schema and one > schama is the same as 1 index in our case, 1 table from the Database. > I read about multicore, but i can not imagine that for each table we want > to use 1 server instance. > do you have a solution for this problem? > > If not, can you tell me where i can find one? > > Thank a lot in advance, > Best regards, > Cynthia Dematteis-Krug > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > -- Lance Norskog [email protected]
