欢迎您访问程序员文章站本站旨在为大家提供分享程序员计算机编程知识!
您现在的位置是: 首页

druid查询

程序员文章站 2022-05-10 18:21:58
...

druid为各种场景提供了丰富的查询类型。 查询由各种JSON属性组合而成,不同类型的查询,JSON属性不同,下面主要介绍常用的查询类型。

1. Components

  1. Datasources

    一个数据源等价于druid表。此外,一个查询也可以作为数据源,提供类似于子查询的功能。查询数据源目前只支持GroupBy查询

    • Table Data Source

      最常用的类型

      1
      2
      3
      4
      5
      6
      7
      8
      9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      {
        "queryType""scan",
        "dataSource": {
          "type""table",
          "name""wikipedia3"
        },
        "resultFormat""list",
        "columns": [
          "page",
          "countryName",
          "cityName",
          "countryIsoCode"
        ],
        "intervals": [
          "2016-06-27/2016-06-28"
        ],
        "limit": 5
      }

        

    • Union Data Source

      联合数据源中的数据源必须拥有相同的schema。联合查询只能被发送给broker/router节点,不支持直接发送到历史节点

      1
      2
      3
      4
      {
         "type""union",
         "dataSources": ["<string_value1>""<string_value2>""<string_value3>", ... ]
      }

        

  2. Filters

    • Selector filter

      等价于sql的where countryIsoCode = 'US'

      1
      2
      3
      4
      5
      6
      7
      8
      9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      {
            "queryType""scan",
            "dataSource": {
              "type""table",
              "name""wikipedia3"
            },
            "filter": {
              "type""selector",
              "dimension""countryIsoCode",
              "value""US"
            },
            "resultFormat""list",
            "columns": [
              "page",
              "countryName",
              "cityName"
            ],
            "intervals": [
              "2016-06-27/2016-06-28"
            ],
            "limit": 5
          }

        

    • Column Comparison filter

      等价于sql的where countryName = cityName

      1
      2
      3
      4
      5
      6
      7
      8
      9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      24
      {
        "queryType""scan",
        "dataSource": {
          "type""table",
          "name""wikipedia3"
        },
        "filter": {
          "type""columnComparison",
          "dimensions": [
            "countryName",
            "cityName"
          ]
        },
        "resultFormat""list",
        "columns": [
          "page",
          "countryName",
          "cityName"
        ],
        "intervals": [
          "2016-06-27/2016-06-28"
        ],
        "limit": 5
      }

        

    • Regular expression filter

      正则表达式,支持标准的java正则表达式,下面的查询表示countryIsoCode以U开头

      1
      2
      3
      4
      5
      6
      7
      8
      9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      {
        "queryType""scan",
        "dataSource": {
          "type""table",
          "name""wikipedia3"
        },
        "filter": {
          "type""regex",
          "dimension""countryIsoCode",
          "pattern""^U"
        },
        "resultFormat""list",
        "columns": [
          "page",
          "countryName",
          "cityName",
          "countryIsoCode"
        ],
        "intervals": [
          "2016-06-27/2016-06-28"
        ],
        "limit": 5
      }

        

    • Logical expression filters

      支持and or not,下面的等价于 where countryIsoCode = 'US' and cityName = 'New York'

      1
      2
      3
      4
      5
      6
      7
      8
      9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      24
      25
      26
      27
      28
      29
      30
      31
      32
      33
      {
            "queryType""scan",
            "dataSource": {
              "type""table",
              "name""wikipedia3"
            },
            "filter": {
              "type""and",
              "fields": [
                {
                  "type""selector",
                  "dimension""countryIsoCode",
                  "value""US"
                },
                {
                  "type""selector",
                  "dimension""cityName",
                  "value""New York"
                }
              ]
            },
            "resultFormat""list",
            "columns": [
              "page",
              "countryName",
              "cityName",
              "countryIsoCode"
            ],
            "intervals": [
              "2016-06-27/2016-06-28"
            ],
            "limit": 5
          }

        

    • JavaScript filter

      1
      2
      3
      4
      5
      6
      7
      8
      9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      {
        "queryType""scan",
        "dataSource": {
          "type""table",
          "name""wikipedia3"
        },
        "filter": {
          "type""javascript",
          "dimension" "countryIsoCode",
          "function" "function(value) { return (value == 'US' || value == 'CN') }"
        },
        "resultFormat""list",
        "columns": [
          "page",
          "countryName",
          "cityName",
          "countryIsoCode"
        ],
        "intervals": [
          "2016-06-27/2016-06-28"
        ],
        "limit": 5
      }

        

    • Search filter

      用于部分字符串匹配,如下面的表示包含foo,并且对大小写不敏感

      1
      2
      3
      4
      5
      6
      7
      8
      9
      10
      {
        "filter": {
          "type""search",
          "dimension""product",
          "query": {
            "type""insensitive_contains",
            "value""foo"
          }
        }
      }

        

    • In filter

      等价于where countryIsoCode in ('US', 'CN')

      1
      2
      3
      4
      5
      6
      7
      8
      9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      {
        "queryType""scan",
        "dataSource": {
          "type""table",
          "name""wikipedia3"
        },
        "filter": {
          "type""in",
          "dimension" "countryIsoCode",
          "values": ["US""CN"]
        },
        "resultFormat""list",
        "columns": [
          "page",
          "countryName",
          "cityName",
          "countryIsoCode"
        ],
        "intervals": [
          "2016-06-27/2016-06-28"
        ],
        "limit": 5
      }

        

    • Like filter

      等价于where countryIsoCode like '%U'

      1
      2
      3
      4
      5
      6
      7
      8
      9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      {
        "queryType""scan",
        "dataSource": {
          "type""table",
          "name""wikipedia3"
        },
        "filter": {
          "type""like",
          "dimension" "countryIsoCode",
          "pattern""%U"
        },
        "resultFormat""list",
        "columns": [
          "page",
          "countryName",
          "cityName",
          "countryIsoCode"
        ],
        "intervals": [
          "2016-06-27/2016-06-28"
        ],
        "limit": 5
      }

        

    • Bound filter

      等价于 "CN" < countryIsoCode < "US"

      1
      2
      3
      4
      5
      6
      7
      8
      9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      24
      25
      26
      27
      28
      {
        "queryType""scan",
        "dataSource": {
          "type""table",
          "name""wikipedia3"
        },
        "filter": {
          "type""bound",
          "dimension" "countryIsoCode",
          "lower""CN",
          "lowerStrict"true,
          "upper""US" ,
          "ordering""numeric",
          "upperStrict"true,
          "ordering""lexicographic"
        },
        "resultFormat""list",
        "columns": [
          "page",
          "countryName",
          "cityName",
          "countryIsoCode"
        ],
        "intervals": [
          "2016-06-27/2016-06-28"
        ],
        "limit": 5
      }

        

  3. Aggregations

    • Count aggregator

      1
      2
      3
      4
      5
      6
      7
      8
      select
          page,
          count(*) as num
      from wikipedia3
      where "__time" BETWEEN TIMESTAMP '2016-06-27 00:00:00' AND TIMESTAMP '2016-06-28 00:00:00'
      group by page
      order by num desc
      limit 5

        

      1
      2
      3
      4
      5
      6
      7
      8
      9
      10
      11
      12
      13
      14
      15
      16
      17
      {
        "queryType""topN",
        "dataSource""wikipedia3",
        "dimension""page",
        "threshold": 5,
        "metric""num",
        "granularity""all",
        "aggregations": [
          {
            "type""count",
            "name""num"
          }
        ],
        "intervals": [
          "2016-06-27/2016-06-28"
        ]
      }

        

    • Sum aggregators

      longSum、doubleSum、floatSum

      1
      2
      3
      4
      5
      6
      7
      8
      select
          page,
          sum(delta) as num
      from wikipedia3
      where "__time" BETWEEN TIMESTAMP '2016-06-27 00:00:00' AND TIMESTAMP '2016-06-28 00:00:00'
      group by page
      order by page asc
      limit 5

        

      1
      2
      3
      4
      5
      6
      7
      8
      9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      {
        "queryType""topN",
        "dataSource""wikipedia3",
        "dimension""page",
        "threshold": 5,
        "metric""num",
        "granularity""all",
        "aggregations": [
          {
            "type""longSum",
            "name""num",
            "fieldName" "delta"
          }
        ],
        "intervals": [
          "2016-06-27/2016-06-28"
        ]
      }

        

    • Min / Max aggregators

      doubleMin、doubleMax、floatMin、floatMax、longMin、longMax

      1
      2
      3
      4
      5
      6
      7
      8
      select
          page,
          max(delta) as num
      from wikipedia3
      where "__time" BETWEEN TIMESTAMP '2016-06-27 00:00:00' AND TIMESTAMP '2016-06-28 00:00:00'
      group by page
      order by page asc
      limit 5

        

      1
      2
      3
      4
      5
      6
      7
      8
      9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      {
        "queryType""topN",
        "dataSource""wikipedia3",
        "dimension""page",
        "threshold": 5,
        "metric""num",
        "granularity""all",
        "aggregations": [
          {
            "type""longMax",
            "name""num",
            "fieldName" "delta"
          }
        ],
        "intervals": [
          "2016-06-27/2016-06-28"
        ]
      }

        

    • First / Last aggregator

      不能在数据摄入的时候使用,只能用于查询

      Last:最大时间戳对应的数据,0 if no row exist;First最小时间戳对应的数据,0 if no row exist

    • JavaScript aggregator

  4. Post Aggregations

    对Aggregations的结果进行二次加工并输出,最终的结果既包含Aggregations的结果也包含Post Aggregations的结果

2. Timeseries

统计一段时间内的汇总数据

1
2
3
4
SELECT count(*) as num,
sum(added)
FROM wikipedia
WHERE "__time" BETWEEN TIMESTAMP '2016-06-27 00:00:00' AND TIMESTAMP '2016-06-27 23:59:59'

  

1
2
3
4
5
6
7
8
9
10
{
  "queryType""timeseries",
  "dataSource""wikipedia3",
  "granularity""all",
  "aggregations": [
    "type""count""name""count" },
    "type""longSum""name""added""fieldName""added" }
  ],
  "intervals": [ "2016-06-27/2016-06-28" ]
}

  

3. TopN

返回前N条数据,并可以按照metric排序,可以支持维度,但只有一个

1
2
3
4
5
6
7
8
SELECT
    page,
    sum(added) as num
FROM wikipedia
WHERE "__time" BETWEEN TIMESTAMP '2016-06-27 00:00:00' AND TIMESTAMP '2016-06-27 23:59:59'
group by page
order by num desc
limit 5

  

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
{
  "queryType""topN",
  "dataSource""wikipedia3",
  "dimension""page",
  "threshold": 5,
  "metric""added",
  "granularity""all",
  "aggregations": [
    {
      "type""doubleSum",
      "name""added",
      "fieldName""added"
    }
  ],
  "intervals": [ "2016-06-27/2016-06-28" ]
}

  

4. GroupBy

能对指定的多个维度分组,也支持对指定的维度排序,也支持limit,但是性能比TopN和Timeseries要差很多

1
2
3
4
5
6
7
8
9
10
SELECT
    page,
    countryName,
    sum(added) as num,
    sum(delta) as num2
FROM wikipedia
WHERE "__time" BETWEEN TIMESTAMP '2016-06-27 00:00:00' AND TIMESTAMP '2016-06-27 23:59:59'
group by page,countryName
order by num desc
limit 5

  

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
{
  "queryType""groupBy",
  "dataSource""wikipedia3",
  "granularity""all",
  "dimensions": [
    "page",
    "countryName"
  ],
  "limitSpec": {
    "type""default",
    "limit": 5,
    "columns": [
      {
        "dimension""added",
        "direction""descending",
        "dimensionOrder": {
          "type""numeric"
        }
      }
    ]
  },
  "aggregations": [
    {
      "type""longSum",
      "name""added",
      "fieldName""added"
    },
    {
      "type""longSum",
      "name""delta",
      "fieldName""delta"
    }
  ],
  "intervals": [
    "2016-06-27/2016-06-28"
  ]
}

  

5. Search

类似于like操作,可以查询多个维度列,不支持聚合

1
2
3
4
5
6
7
SELECT
page,
countryName
FROM wikipedia
WHERE "__time" BETWEEN TIMESTAMP '2016-06-27 00:00:00' AND TIMESTAMP '2016-06-27 23:59:59'
and page like '%C' or countryName like '%C'
limit 5

  

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
{
  "queryType""search",
  "dataSource""wikipedia3",
  "granularity""all",
  "dimensions": [
    "page",
    "countryName"
  ],
  "query": {
    "type""insensitive_contains",
    "value""C"
  },
  "sort" : {
    "type""lexicographic"
  },
  "limit": 5,
  "intervals": [
    "2016-06-27/2016-06-28"
  ]
}

  

6. Select

查数据,不支持聚合,但支持分页,排序

1
2
3
4
5
SELECT
*
FROM wikipedia
WHERE "__time" BETWEEN TIMESTAMP '2016-06-27 00:00:00' AND TIMESTAMP '2016-06-27 23:59:59'
limit 0,5

  

1
2
3
4
5
6
7
8
9
10
11
{
  "queryType""select",
  "dataSource""wikipedia3",
  "granularity""all",
  "dimensions":[],
  "metrics":[],
  "pagingSpec":{"pagingIdentifiers": {}, "threshold":5},
  "intervals": [
    "2016-06-27/2016-06-28"
  ]
}

  

7. Scan

类似于Select,但不支持分页,但是如果没有分页需求,推荐使用这个,性能比Select好

1
2
3
4
5
SELECT
page,countryName
FROM wikipedia
WHERE "__time" BETWEEN TIMESTAMP '2016-06-27 00:00:00' AND TIMESTAMP '2016-06-27 23:59:59'
limit 5

  

1
2
3
4
5
6
7
8
9
10
11
{
   "queryType""scan",
   "dataSource""wikipedia3",
   "resultFormat""list",
   "columns":["page","countryName"],
   "intervals": [
     "2016-06-27/2016-06-28"
   ],
   "batchSize":20480,
   "limit":5
 }
相关标签: druid