How to use ingest DeDot processor

Elastic Stack Elasticsearch
1

We want to replace dots with underscores while ingesting the data. I found de-dot processor which seems to be a good solution. Here is the PR that was merged early last year, but I don't find that in the code anymore or in ES documentation.

Here is the PR: https://github.com/elastic/elasticsearch/pull/15968/files

Could someone please help answering the following questions:

  1. Is this supported.
  2. If the answer to question 1 is yes, how do I use dedot processor?
  3. What is the impact on write throughput because of this processor? Has there been any tests conducted?

Look forward to hearing from you.

2

Is it possible that you are referring to the dot_expander processor, see also the corresponding [PR] (https://github.com/elastic/elasticsearch/pull/20078)

--Alex

3

Most likely not. Two reasons below:

  1. Dot expander seems to require the field to expand. We do not know the fieldname.
  2. Dot expander wouldn't handle cases like "a....b.c" and ES would simply return an exception.

Dedot would simply replace the dots with a separator of your choice e.g. underscore. DeDot processor was added for the same reason, but the recent version of ES (5.3) does not have that processor in the code. I added the code myself and I confirmed that it is working, but I am not sure about the performance impact and / or the reason for removing it in the first place.

4

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