I'm trying to get some basic working by setting the _score of a document, using function scores and/or painless scripts. I plan to use the basic block to find the mechanism to build out my final use case. Sadly it already fails. My data looks like this:
{
    "brand": "Apple",
    "flavourScore": 1.1,
    "price": 6.5
}
I have 9 items (3 Apples, 3 Bananas, 3 Tomatoes, differing in flavourScore and price). I've created 1 mapping for the index where they are in (couldn't post messages otherwise).
All documents have a _score of 1 by default. Mildly strange, but as I want to recompute the _score I don't care that much. I had tried a few examples found here and there, adapted them, but there is little in the form of debugging you can do. So I decided ny first test would be to just reset all the _score values to a different constant (as they are all 1 now as well). However.....nothing gets updated on ES64 (Docker image). Any clues why the GET on the _search endpoint for this index fails:
{
  "query": {
    "function_score": {
      "query": {
        "match_all": {}
      },
      "functions": [
        {
          "script_score": {
            "script": {
              "lang": "painless",
              "inline": "_score = 0.18;"
            }
          }
        }
      ],
      "boost_mode": "replace"
    }
  },
  "sort": [
    "_score"
  ]
}
I'd expect all _score values to be 0.18 when I do a match_all query, but they stay 1.
Any insights, examples, clues are highly appreciated - especially as my next step will to have the script compute the value for _score
Thanks,
Maarten
