Can not make aggregations to work on string field (ES 5.5.1)


(Zoltán Zvara) #1

Dear Community,

I have made a small fun PoC last year, 2016. December 18th, on which I inserted tweets (Twitter4J) into ES, and visualized with Kibana.

image

The above visualization shows significant terms that appear in tweets that contain "trump", and on the Y axis, their corresponding sentiment value. I prepared no mappings by hand, fields have been indexed dynamically and appeared in Kibana - surprisingly and fortunately.

Now, using the same data, I would like to define my own mappings, but could not produce the same visualization. The "tweet.text" if analyzed, no aggregations are available. If I create a "raw" field mapping as well, it is aggregatable, but definitely not going to produce the same results. I also added the following to the field definition:

"term_vector": "with_positions_offsets_payloads",
"index_options": "offsets"

I'm using the "standard" analyzer, and field type "text".

I kindly as the Community to help me to achieve such a visualization on any text field.

Cheers,
Zoltán


(Zoltán Zvara) #2

It seems to me that aggregations on analyzed string fields have been removed. The following is a working example from Kibana 5.0.0-alpha4.

This can not be done in Kibana 5.5.1.

I would be happy to receive a conformation from the Community, that this functionality has been removed from Kibana.


(Barnabás Balázs) #3

This feature would be nice to have for my application too, I hope this issue will be fixed soon.


(system) #4

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.