0

We've run into a problem with our Datastax Enterprise search workload. Our version is [cqlsh 5.0.1 | Cassandra 2.1.9.791 | DSE 4.8.0 | CQL spec 3.2.0 | Native protocol v3]. We have a cassandra schema for this table with a map collection properties map<text, text> and in the solr schema.xml we have <dynamicField name="properties*" type="text" indexed="true" stored="true" />.

The problem is that we've inserted data incorrectly, without using the prefix as described here. Is there a way to make this data searchable without losing all of the data that's already in the table?

LHWizard
  • 2,121
  • 19
  • 30

1 Answers1

1

I wrote a whole blog post about it:

http://www.sestevez.com/brians-loader-unloader-for-search/

TL;DR you have to re-insert the data, but the post has a one liner on how to do it with cassandra-loader and awk.

phact
  • 7,305
  • 23
  • 27
  • In your blog, you warn about using collections for large amounts of data. How large is "large"? With `nodetool cfstats` I see `Space used (total): 83112983` which doesn't seem too large. – LHWizard Apr 05 '16 at 19:26
  • as I think about it, what are some reasonable alternatives, avoiding map collections, to accomplish the same thing? – LHWizard Apr 05 '16 at 20:01
  • Large as in too many vales in a map. Not total size. I'd aim for 10s – phact Apr 05 '16 at 23:09