bq: This started when I turned on "docvalues" You _must_ re-index completely when changing something like this, so the notion of removing the index completely isn't really any extra work.
Here's what I'd do. 1> just create a new collection with your current schema definition and index to _that_. That'll guarantee you don't have anything pre-existing that pollutes your index. 2> verify that this does what you want. Perhaps use a smaller set of docs than your entire corpus. 3> delete your original collection 4> If you require the same name, you can use collection aliasing to make this change transparent. Creating/deleting collections and using collection aliasing are all through the Collections API. Best, Erick On Fri, Jul 24, 2015 at 10:16 AM, shamik <sham...@gmail.com> wrote: > I didn't use the REST API, instead updated the schema manually. > > Can you be specific on removing the data directory content ? I certainly > don't want to wipe out the index. I've four Solr instances, 2 shards with a > replica each. Are you suggesting clearing the index and re-indexing from > scratch ? > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/Unexpected-docvalues-type-error-using-result-grouping-Use-UninvertingReader-or-index-with-docvalues-tp4218939p4219089.html > Sent from the Solr - User mailing list archive at Nabble.com.
