How to get back the facets when passing facet.limit=-1

Question

question

rkagan_146583 asked · Nov 07, 2019 at 12:32 AM

How to get back the facets when passing facet.limit=-1

Hi:

We just upgraded from DSE 5.1.3 to 5.1.7 and right away we lost the facets, that were returning from Solr. This is due to the fact that we sent facet.limit=-1 (unbounded facets). We had to do that due to the fact that we need to get ALL facets and show them on the site, instead of setting some arbitrary upper limit. However DataStax engineers put the breaking change that traps such setting and no facets returned. What should we send to DSE instead to get all facets (while still staying in version 5.1.7). Really appreciate for quick response.

dse dse search solr

2

10 |1000 characters needed characters left characters exceeded

Attachments: Up to 8 attachments (including images) can be used with a maximum of 1.0 MiB each and 10.0 MiB total.

harikrishnan.nair_59682 · Jan 20, 2020 at 11:12 PM

Hi, Would you able to find any solution to the facet limitation you mentioned here ? We do have similar use case, and we are in the process of upgrading to 5.1..17. Wondering if there is any other work around other changing the Jim.options as mentioned in https://docs.datastax.com/en/upgrade/doc/upgrade/datastax_enterprise/upgdDseMinor51x.html

0 ·

Erick Ramirez ♦♦ · Jan 21, 2020 at 08:54 AM

Apologies for having missed this question and that you didn't get an answer. Drafting one now. Cheers!

0 ·

Answer 1 · 2020-01-21T09:34:55Z

Erick Ramirez answered · Jan 21, 2020 at 09:34 AM

@rkagan_146583 The unbounded facets was recently disallowed because we were seeing too many clusters getting unnecessarily overloaded when queries try to pull back tens of thousands of facets. It's an indication that maybe Solr facets are being used in a way they weren't intended for.

In DSE, the maximum number has been capped to 20K. This is an artificial limit to try and prevent queries from bringing nodes down. If you don't specify facet.limit=-1, your queries should return up to 20K facets instead of none.

@harikrishnan.nair_59682 We do not recommend that you override the default limit in the JVM options. We know that if your queries require more than 20K facets (still an excessively high number) your cluster's performance will be severely degraded. Cheers!

2 Share

10 |1000 characters needed characters left characters exceeded

Attachments: Up to 8 attachments (including images) can be used with a maximum of 1.0 MiB each and 10.0 MiB total.

harikrishnan.nair_59682 · Jan 21, 2020 at 09:26 PM

Thanks @Erick Ramirez , Primarily we use faceting to pull the distinct values using dse search query. Is there any option to achieve this ?

0 ·

Erick Ramirez ♦♦ harikrishnan.nair_59682 · Jan 21, 2020 at 09:58 PM

If you have an OLAP use case, you should use Spark in DSE Analytics. Cheers!

0 ·

question