Absolutely, using Mikhail's code would be the first thing I would do. You can see the details in SOLR-6234 and https://github.com/m-khl/lucene-join-solr-query-parser
Otherwise, the only alternative I can think of (without reindexing) would be to run the select query as provided, returning the ID's of the offending documents (for reference, you could use grouping or the collapsing query parser if you just want to get distinct values; faceting is also an option). Then write a script to iterate through a batch of ID's at a time and send a delete to Solr, such as <delete><query>id:(100 OR 101 OR ...)</query></delete>. Since there are many documents to delete, you would want to hold off committing until the end. On 11 October 2014 02:34, Mikhail Khludnev <mkhlud...@griddynamics.com> wrote: > On Fri, Oct 10, 2014 at 6:16 AM, Matthew Nigl <matthew.n...@gmail.com> > wrote: > > > But I get the same response as in > > https://issues.apache.org/jira/browse/SOLR-6357 > > > > there is a mention for cure (SOLR-6234 > <https://issues.apache.org/jira/browse/SOLR-6234>) over there > > -- > Sincerely yours > Mikhail Khludnev > Principal Engineer, > Grid Dynamics > > <http://www.griddynamics.com> > <mkhlud...@griddynamics.com> >