fix 'buckets' value may be a dict

[ivanli1990]

What type of PR is this? (check all applicable)

• Bug Fix

Description

The 'buckets' object in elasticsearch aggregations may be not a list but a dict

Related Tickets & Documents

#2789

Mobile & Desktop Screenshots/Recordings (if there are UI changes)

[lkm]

Great thanks @ivanli1990! Has this got to do with using numbered instead of named aggregation names?Maybe we should add a test case for this result?

[ivanli1990]

Great thanks @ivanli1990! Has this got to do with using numbered instead of named aggregation names?Maybe we should add a test case for this result?

Numbered aggregation name is no matter, but there still some issue with multiple buckets result:

{
  "took": 3,
  "timed_out": false,
  "_shards": {
    "total": 5,
    "successful": 5,
    "skipped": 0,
    "failed": 0
  },
  "hits": {
    "total": 47,
    "max_score": 0,
    "hits": []
  },
  "aggregations": {
    "3": {
      "doc_count_error_upper_bound": 0,
      "sum_other_doc_count": 0,
      "buckets": [
        {
          "1": {
            "value": 15
          },
          "4": {
            "buckets": {
              "state:已关闭": {
                "1": {
                  "value": 11
                },
                "5": {
                  "doc_count_error_upper_bound": 0,
                  "sum_other_doc_count": 0,
                  "buckets": [
                    {
                      "1": {
                        "value": 10
                      },
                      "key": "数据自身问题",
                      "doc_count": 10
                    },
                    {
                      "1": {
                        "value": 1
                      },
                      "key": "未填写原因",
                      "doc_count": 1
                    }
                  ]
                },
                "doc_count": 11
              }
            }
          },
          "key": "数据",
          "doc_count": 15
        },
        {
          "1": {
            "value": 9
          },
          "4": {
            "buckets": {
              "state:已关闭": {
                "1": {
                  "value": 8
                },
                "5": {
                  "doc_count_error_upper_bound": 0,
                  "sum_other_doc_count": 0,
                  "buckets": [
                    {
                      "1": {
                        "value": 8
                      },
                      "key": "未填写原因",
                      "doc_count": 8
                    }
                  ]
                },
                "doc_count": 8
              }
            }
          },
          "key": "多码后端",
          "doc_count": 32
        }
      ]
    }
  }
}

[lkm]

Alright if I could get you to please finish the test (there are no assertions now, it's fine that it would be failing), then I'll try to dig into it again

[ivanli1990]

Alright if I could get you to please finish the test (there are no assertions now, it's fine that it would be failing), then I'll try to dig into it again

hi lkm,
I added 2 cases in test, and I wrote the expect rows and columns in each test case, it's just the result I expected, sort of each row in rows(same in column in columns) maybe different from the actual rows and columns which parse with query result. So, you could just follow the expect rows and columns as a reference.
In addtion, I think 'doc_count_error_upper_bound' and 'sum_other_doc_count' fileds in elasticsearch query result should not be display on redash, user not care these fileds, so they should not add in rows and columns.