score and explanation for dfs queries don't match

[brwe]

Tested 1.7 vs 2.1. The following example:

POST test/doc/1
{
  "text": "Score and explanation should match I think?"
}

POST test/doc/2
{
  "text": "Score and explanation did match in 1.7..."
}


GET test/_search?search_type=dfs_query_then_fetch&explain
{
  "query": {
    "match": {
      "text": "score"
    }
  }
}

for 1.7.0 returns:

{
   "took": 48,
   "timed_out": false,
   "_shards": {
      "total": 5,
      "successful": 5,
      "failed": 0
   },
   "hits": {
      "total": 2,
      "max_score": 0.22295055,
      "hits": [
         {
            "_shard": 2,
            "_node": "5IoTIXdmROKObfLRmqCQYQ",
            "_index": "test",
            "_type": "doc",
            "_id": "1",
            "_score": 0.22295055,
            "_source": {
               "text": "Score and explanation should match I think?"
            },
            "_explanation": {
               "value": 0.22295056,
               "description": "weight(text:score in 0) [PerFieldSimilarity], result of:",
               "details": [
                  {
                     "value": 0.22295056,
                     "description": "score(doc=0,freq=1.0), product of:",
                     "details": ```
         {
            "_shard": 2,
            "_node": "8AbQaexnTwKbvcL1Va4yfw",
            "_index": "test",
            "_type": "doc",
            "_id": "1",
            "_score": 0.22295055,
            "_source": {
               "text": "Score and explanation should match I think?"
            },
            "_explanation": {
               "value": 0.22295056,
               "description": "weight(text:score in 0) [PerFieldSimilarity], result of:",[
                        {
                           "value": 0.99999994,
                           "description": "queryWeight, product of:",
                           "details": [
                              {
                                 "value": 0.5945349,
                                 "description": "idf(docFreq=2, maxDocs=2)"
                              },
                              {
                                 "value": 1.681987,
                                 "description": "queryNorm"
                              }
                           ]
                        },
                        {
                           "value": 0.22295058,
                           "description": "fieldWeight in 0, product of:",
                           "details": [
                              {
                                 "value": 1,
                                 "description": "tf(freq=1.0), with freq of:",
                                 "details": [
                                    {
                                       "value": 1,
                                       "description": "termFreq=1.0"
                                    }
                                 ]
                              },
                              {
                                 "value": 0.5945349,
                                 "description": "idf(docFreq=2, maxDocs=2)"
                              },
                              {
                                 "value": 0.375,
                                 "description": "fieldNorm(doc=0)"
                              }
                           ]
                        }
                     ]
                  }
               ]
            }
         },
         {
            "_shard": 3,
            "_node": "5IoTIXdmROKObfLRmqCQYQ",
            "_index": "test",
            "_type": "doc",
            "_id": "2",
            "_score": 0.22295055,
            "_source": {
               "text": "Score and explanation did match in 1.7..."
            },
            "_explanation": {
               "value": 0.22295056,
               "description": "weight(text:score in 0) [PerFieldSimilarity], result of:",
               "details": [
                  {
                     "value": 0.22295056,
                     "description": "score(doc=0,freq=1.0), product of:",
                     "details": [
                        {
                           "value": 0.99999994,
                           "description": "queryWeight, product of:",
                           "details": [
                              {
                                 "value": 0.5945349,
                                 "description": "idf(docFreq=2, maxDocs=2)"
                              },
                              {
                                 "value": 1.681987,
                                 "description": "queryNorm"
                              }
                           ]
                        },
                        {
                           "value": 0.22295058,
                           "description": "fieldWeight in 0, product of:",
                           "details": [
                              {
                                 "value": 1,
                                 "description": "tf(freq=1.0), with freq of:",
                                 "details": [
                                    {
                                       "value": 1,
                                       "description": "termFreq=1.0"
                                    }
                                 ]
                              },
                              {
                                 "value": 0.5945349,
                                 "description": "idf(docFreq=2, maxDocs=2)"
                              },
                              {
                                 "value": 0.375,
                                 "description": "fieldNorm(doc=0)"
                              }
                           ]
                        }
                     ]
                  }
               ]
            }
         }
      ]
   }
}

and in 2.1:

{
   "took": 46,
   "timed_out": false,
   "_shards": {
      "total": 5,
      "successful": 5,
      "failed": 0
   },
   "hits": {
      "total": 2,
      "max_score": 0.22295055,
      "hits": [
         {
            "_shard": 2,
            "_node": "hcC0CtAfTSSILbEq32T_fw",
            "_index": "test",
            "_type": "doc",
            "_id": "2",
            "_score": 0.22295055,
            "_source": {
               "text": "Score and explanation did match in 1.7..."
            },
            "_explanation": {
               "value": 0.11506981,
               "description": "weight(text:score in 0) [PerFieldSimilarity], result of:",
               "details": [
                  {
                     "value": 0.11506981,
                     "description": "fieldWeight in 0, product of:",
                     "details": [
                        {
                           "value": 1,
                           "description": "tf(freq=1.0), with freq of:",
                           "details": [
                              {
                                 "value": 1,
                                 "description": "termFreq=1.0",
                                 "details": []
                              }
                           ]
                        },
                        {
                           "value": 0.30685282,
                           "description": "idf(docFreq=1, maxDocs=1)",
                           "details": []
                        },
                        {
                           "value": 0.375,
                           "description": "fieldNorm(doc=0)",
                           "details": []
                        }
                     ]
                  }
               ]
            }
         },
         {
            "_shard": 3,
            "_node": "hcC0CtAfTSSILbEq32T_fw",
            "_index": "test",
            "_type": "doc",
            "_id": "1",
            "_score": 0.22295055,
            "_source": {
               "text": "Score and explanation should match I think?"
            },
            "_explanation": {
               "value": 0.11506981,
               "description": "weight(text:score in 0) [PerFieldSimilarity], result of:",
               "details": [
                  {
                     "value": 0.11506981,
                     "description": "fieldWeight in 0, product of:",
                     "details": [
                        {
                           "value": 1,
                           "description": "tf(freq=1.0), with freq of:",
                           "details": [
                              {
                                 "value": 1,
                                 "description": "termFreq=1.0",
                                 "details": []
                              }
                           ]
                        },
                        {
                           "value": 0.30685282,
                           "description": "idf(docFreq=1, maxDocs=1)",
                           "details": []
                        },
                        {
                           "value": 0.375,
                           "description": "fieldNorm(doc=0)",
                           "details": []
                        }
                     ]
                  }
               ]
            }
         }
      ]
   }
}

[rjernst]

@brwe Can you explain what "doesn't match"? The explanations and scores look the same to me, but maybe I'm just not seeing it in the page-of-json. :)

[jpountz]

@rjernst I wondered the same but I think I see it now: in the 2.1 output _score is different from _explanation.value.

[clintongormley]

Related to #2612?

[sylvinus]

I have the same issue. Any clue on how to fix or a workaround?

[gmoskovicz]

I can confirm that this still happens in 2.3.

Looks like when using dfs, the explain ignores the fact that we need to merge the results from the information of the shards. The explain value when using dfs is the same score that you get if you do not use dfs, but instead just query then fetch.

[gmoskovicz]

        "_score": 0.2202036,     <---------- dfs score
        "_source": {
          "city": "montevideo",
          "cityAliases": "mvd"
        },
        "_explanation": {
          "value": 0.2417773,       <----------- the score for non-dfg
          "description": "max of:",
          "details": [
            {
              "value": 0.2417773,

[gmoskovicz]

Are there any test cases that verifies that the explanation root value is equals to the hit score?