Solr 1.3.0 Released
Apache Solr 1.3.0 has been released. This version contains many, many improvements and bug fixes. High on my list are things like a good first step on distributed search support, integrated spell checking, support for Lucene’s “More Like This”, and the much needed Data Import Handler. Of course, one can’t forget about the numerous performance improvements. Since this release is using a very recent version of Lucene, it takes advantage of all of the improvements in Lucene 2.3 and above, which are quite significant.
With the Data Import Handler, for example, one can now easily hook Solr up to any database that supports JDBC and have it index the database for search. Since most databases aren’t particularly good at text search, this will be a really attractive option for a lot of people who want to index a database.
The new distributed search support is also quite nice in that it allows developers to work with even larger collections without having to roll their own complex solution in low level Lucene. While it currently requires some operations support to be bullet proof (think load balancers, etc.), it is a great first step and I know it is already used in production by several large installations.
As for spell checking, it used to be that one had to issue a separate query to get spelling results through the SpellcheckRequestHandler, now, thanks to Solr’s pluggable SearchComponent architecture, Google-style “Did you mean” results are available in the response with a query (assuming you turn the component on).
We also are now publishing maven artifacts (but note, I forgot to do it w/ the original distribution process, so they may not show up yet).
For all the features, fixes, etc. see the announcement on the Solr homepage and the link to the release notes.
