Performance of Range queries on Keyword fields

Elastic Stack Elasticsearch
1

Hello, how is the performance of Range queries on Keyword (string) fields?

The date fields with values in ISO formats [yyyy-MM-dd'T'HH:mm:ss. SSSXXX] are indexed as keyword instead of date.
There is a need to filter the records with date range.

The record count per shard in Prod cluster indexes is a couple of lacs.
Elasticsearch version: 7.10.2

Is it very inefficient compared to the date and long fields? Should we re-index the data with date mappings?

Thanks

2

Welcome!

You can not really use keyword type for date comparison. You'd better to reindex.

While you are at it, please upgrade as your version is too old.

4

Thanks.
Range query works on keyword fields with a query like below

{
  "query": {
    "range": {
      "fieldName": {
        "gte": "valueA",
        "lte": "valueB"
      }
    }
  }
}

Is there any documentation/references related to its performance?

5

My guess is that it's faster with optimized data structures. Here date data type for dates.