According to the documentation, when doing a regex search in opensearch, the regex is applied to the individual tokens in the field, not the entire field.

Because of this, a query like

  {
    "query": {
      "regexp": {
        "myfield": ".+ABC.+"
      }
    }
  }

does not return any results when the term “ABC” is preceeded by whitespace.

So for example, the field myfield has a string that has spaces (and therefore separate terms):

The fields FIELD1 and FIELD2 are different: ABC4345 vs 2405987390

Searching for this exact message (with “matches”) will return it, but how, within the context of opensearch, could a query be written to find strings like .+are different.+ABC.+vs.+ or even just .+ABC.+ ?