thanks Alex for taking out your valuable time and helping us to understand
better.
Cheers!
Kshitij
On Mon, Apr 10, 2017 at 4:00 PM, alessandro.benedetti
wrote:
> It really depends on the schema change...
> Any addition/deletion usually implies you can avoid re-indexing if you
> don't
> care the
It really depends on the schema change...
Any addition/deletion usually implies you can avoid re-indexing if you don't
care the old documents will remain outdated.
But doing a type change, or a change to the data structures involved ( such
enabling docValues, norms ect ect) without a full re-index
Hi Alex,
After full re-indexing things work out fine.
But is there any other way to make schema changes on the go?
Or we have to reindex entire data whenever a schema change is done?
we are having 30-40 million documents and it is a tedious and time taking
task.
What other approaches are there
Did you do a full reindex? Try completely deleting the index and
redoing it from scratch (at least as a test). If you have left over
documents and changed type definitions, things may get messy. If
that's too hard, just index a single record into a separate collection
with matching-definition and c
Hi Alex,
Thanks for the response.
1. Actually everything was working fine earlier unless I made changes on a
dynamic field whose facets are being created, I changed the field type to
"strings" earlier I was tokenizing my field on based delimeter '_'.
2. When I made the changes and started indexi
What happens when you facet on a 'all document query' (q=*:*)? Are you
sure your facet fields actually have the content? If they are stored,
they should be returned with the query. If they are not stored, you
could see what they contain in the Schema screen of the Admin UI (load
tokens checkbox).
