为什么Elasticsearch的字段数限制为1000

Question

1. Why has Elasticsearch put a limit of 1000 on the total number of fields?
2. Will it cause any degradation in performance?
3. If it will, why? There is no such limit in lucene.

Also, can someone please explain these settings more in detail.

Answer 1

I believe the explanation from the documentation itself is pretty good when it comes to these settings.

There are situation where the mapping of indices is dynamic (logging use cases usually or some rogue documents generator in the users' infrastructure) and having it like this could lead to hundreds of thousands of fields in indices. When this usually happens, only a handful of documents have values in those gazillion fields, while most of the documents don't. This, from the disk usage point of view, is not a great thing. Why this could be a problem, again is very well explained in another section of the documentation .

The bad part, though, is that most of the times the users didn't plan or didn't expect to have such a high number of fields and the soft limit we introduced for fields count is to protect users in these scenarios from hurting themselves unintentionally.

If you have a well thought mapping for your indices and you know those 1000 fields limit is not for you, you can change it to a number that fits your use case and scenario:

PUT /test/_settings { "index.mapping.total_fields.limit": 2000 }