Hive 差集实现

1. IN query plan

1. NOT IN

select * from id_full as a
where a.id not in (select id from id_incr where day=20180512)
STAGE DEPENDENCIES:
  Stage-4 is a root stage
  Stage-8 depends on stages: Stage-4 , consists of Stage-10, Stage-1
  Stage-10 has a backup stage: Stage-1
  Stage-7 depends on stages: Stage-10
  Stage-6 depends on stages: Stage-1, Stage-7 , consists of Stage-9, Stage-2
  Stage-9 has a backup stage: Stage-2
  Stage-5 depends on stages: Stage-9
  Stage-2
  Stage-1
  Stage-0 depends on stages: Stage-5, Stage-2

主要执行job的有以下3个, stage详细信息:

stage-4

Stage: Stage-4
  Map Reduce
    Map Operator Tree:
        TableScan
          alias: id_incr
          filterExpr: ((day = 20180512) and id is null) (type: boolean)
          Statistics: Num rows: 14933063 Data size: 15054692352 Basic stats: COMPLETE Column stats: NONE
          Filter Operator
            predicate: id is null (type: boolean)
            Statistics: Num rows: 7466531 Data size: 7527345671 Basic stats: COMPLETE Column stats: NONE
            Select Operator
              Statistics: Num rows: 7466531 Data size: 7527345671 Basic stats: COMPLETE Column stats: NONE
              Group By Operator
                aggregations: count()
                mode: hash
                outputColumnNames: _col0
                Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: NONE
                Reduce Output Operator
                  sort order: 
                  Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: NONE
                  value expressions: _col0 (type: bigint)
    Reduce Operator Tree:
      Group By Operator
        aggregations: count(VALUE._col0)
        mode: mergepartial
        outputColumnNames: _col0
        Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: NONE
        Filter Operator
          predicate: (_col0 = 0) (type: boolean)
          Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: NONE
          Select Operator
            expressions: 0 (type: bigint)
            outputColumnNames: _col0
            Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: NONE
            Group By Operator
              keys: _col0 (type: bigint)
              mode: hash
              outputColumnNames: _col0
              Statistics: Num rows: 1 Data size: 8 Basic stats: COMPLETE Column stats: NONE
              File Output Operator
                compressed: false
                table:
                    input format: org.apache.hadoop.mapred.SequenceFileInputFormat
                    output format: org.apache.hadoop.hive.ql.io.HiveSequenceFileOutputFormat
                    serde: org.apache.hadoop.hive.serde2.lazybinary.LazyBinarySerDe

不太明白这个stage的意义,根据这个自己实现了一个sql可以给出类似的stage

select * from (
  select count(id) as cnt from id_incr where day=20180512 and id is null
) as a
where cnt = 0

stage-7

Stage: Stage-7
  Map Reduce
    Map Operator Tree:
        TableScan
          alias: a
          Statistics: Num rows: 4025608842 Data size: 8412504677023 Basic stats: COMPLETE Column stats: NONE
          Map Join Operator
            condition map:
                 Left Semi Join 0 to 1
            keys:
              0 
              1 
            outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9, _col10, _col11, _col12, _col13, _col14, _col15, _col16, _col17, _col18, _col19
            Statistics: Num rows: 4428169822 Data size: 9253755345295 Basic stats: COMPLETE Column stats: NONE
            File Output Operator
              compressed: false
              table:
                  input format: org.apache.hadoop.mapred.SequenceFileInputFormat
                  output format: org.apache.hadoop.hive.ql.io.HiveSequenceFileOutputFormat
                  serde: org.apache.hadoop.hive.serde2.lazybinary.LazyBinarySerDe
    Local Work:
      Map Reduce Local Work

有个mapjoin的动作,没有reduce

Stage-2

Stage: Stage-2
    Map Reduce
      Map Operator Tree:
          TableScan
            Reduce Output Operator
              key expressions: _col0 (type: string)
              sort order: +
              Map-reduce partition columns: _col0 (type: string)
              Statistics: Num rows: 4428169822 Data size: 9253755345295 Basic stats: COMPLETE Column stats: NONE
              value expressions: _col1 (type: string), _col2 (type: string), _col3 (type: string), _col4 (type: string), _col5 (type: string), _col6 (type: string), _col7 (type: string), _col8 (type: string), _col9 (type: string), _col10 (type: string), _col11 (type: string), _col12 (type: string), _col13 (type: string), _col14 (type: string), _col15 (type: string), _col16 (type: string), _col17 (type: string), _col18 (type: string), _col19 (type: string)
          TableScan
            alias: id_incr
            filterExpr: (day = 20180512) (type: boolean)
            Statistics: Num rows: 14933063 Data size: 15054692244 Basic stats: COMPLETE Column stats: NONE
            Select Operator
              expressions: id (type: string)
              outputColumnNames: _col0
              Statistics: Num rows: 14933063 Data size: 15054692244 Basic stats: COMPLETE Column stats: NONE
              Reduce Output Operator
                key expressions: _col0 (type: string)
                sort order: +
                Map-reduce partition columns: _col0 (type: string)
                Statistics: Num rows: 14933063 Data size: 15054692244 Basic stats: COMPLETE Column stats: NONE
      Reduce Operator Tree:
        Join Operator
          condition map:
               Left Outer Join0 to 1
          keys:
            0 _col0 (type: string)
            1 _col0 (type: string)
          outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9, _col10, _col11, _col12, _col13, _col14, _col15, _col16, _col17, _col18, _col19, _col24
          Statistics: Num rows: 4870986909 Data size: 10179131100451 Basic stats: COMPLETE Column stats: NONE
          Filter Operator
            predicate: _col24 is null (type: boolean)
            Statistics: Num rows: 2435493454 Data size: 5089565549180 Basic stats: COMPLETE Column stats: NONE
            Select Operator
              expressions: _col0 (type: string), _col1 (type: string), _col2 (type: string), _col3 (type: string), _col4 (type: string), _col5 (type: string), _col6 (type: string), _col7 (type: string), _col8 (type: string), _col9 (type: string), _col10 (type: string), _col11 (type: string), _col12 (type: string), _col13 (type: string), _col14 (type: string), _col15 (type: string), _col16 (type: string), _col17 (type: string), _col18 (type: string), _col19 (type: string)
              outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9, _col10, _col11, _col12, _col13, _col14, _col15, _col16, _col17, _col18, _col19
              Statistics: Num rows: 2435493454 Data size: 5089565549180 Basic stats: COMPLETE Column stats: NONE
              File Output Operator
                compressed: false
                Statistics: Num rows: 2435493454 Data size: 5089565549180 Basic stats: COMPLETE Column stats: NONE
                table:
                    input format: org.apache.hadoop.mapred.TextInputFormat
                    output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat
                    serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe

map: 两个TableScan读取数据

reduce: 两表join,然后再predicate: _col24 is null (type: boolean)

2. IN examples

select * from (
    select stack(4, 
        'a', null, 'b', 'a'
    ) as key
) a
where a.key not in  (select key from (select stack(2, 'a', null) as key) b);

output:

# nothing
select * from (
    select stack(4, 
        'a', null, 'b', 'a'
    ) as key
) a
where a.key not in  ('a', null);

output:

# nothing

IN语句中只要有null就不会查出数据

3. EXISTS query plan

1. NOT EXISTS

select * from id_full as a
where not exists (select id from id_incr as b where day=20180512 and a.id =b.id)
STAGE DEPENDENCIES:
  Stage-4 is a root stage , consists of Stage-1
  Stage-1
  Stage-0 depends on stages: Stage-1

STAGE PLANS:
  Stage: Stage-4
    Conditional Operator

  Stage: Stage-1
    Map Reduce
      Map Operator Tree:
          TableScan
            alias: a
            Statistics: Num rows: 4025608842 Data size: 8412504677023 Basic stats: COMPLETE Column stats: NONE
            Reduce Output Operator
              key expressions: id (type: string)
              sort order: +
              Map-reduce partition columns: id (type: string)
              Statistics: Num rows: 4025608842 Data size: 8412504677023 Basic stats: COMPLETE Column stats: NONE
              value expressions: _col1 (type: string), ....
          TableScan
            alias: b
            filterExpr: (day = 20180512) (type: boolean)
            Statistics: Num rows: 14933063 Data size: 15054692244 Basic stats: COMPLETE Column stats: NONE
            Select Operator
              expressions: id (type: string)
              outputColumnNames: _col1
              Statistics: Num rows: 14933063 Data size: 15054692244 Basic stats: COMPLETE Column stats: NONE
              Reduce Output Operator
                key expressions: _col1 (type: string)
                sort order: +
                Map-reduce partition columns: _col1 (type: string)
                Statistics: Num rows: 14933063 Data size: 15054692244 Basic stats: COMPLETE Column stats: NONE
      Reduce Operator Tree:
        Join Operator
          condition map:
               Left Outer Join0 to 1
          keys:
            0 id (type: string)
            1 _col1 (type: string)
          outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9, _col10, _col11, _col12, _col13, _col14, _col15, _col16, _col17, _col18, _col19, _col25
          Statistics: Num rows: 4428169822 Data size: 9253755345295 Basic stats: COMPLETE Column stats: NONE
          Filter Operator
            predicate: _col25 is null (type: boolean)
            Statistics: Num rows: 2214084911 Data size: 4626877672647 Basic stats: COMPLETE Column stats: NONE
            Select Operator
              expressions: _col0 (type: string), _col1 (type: string), _col2 (type: string), _col3 (type: string), _col4 (type: string), _col5 (type: string), _col6 (type: string), _col7 (type: string), _col8 (type: string), _col9 (type: string), _col10 (type: string), _col11 (type: string), _col12 (type: string), _col13 (type: string), _col14 (type: string), _col15 (type: string), _col16 (type: string), _col17 (type: string), _col18 (type: string), _col19 (type: string)
              outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9, _col10, _col11, _col12, _col13, _col14, _col15, _col16, _col17, _col18, _col19
              Statistics: Num rows: 2214084911 Data size: 4626877672647 Basic stats: COMPLETE Column stats: NONE
              File Output Operator
                compressed: false
                Statistics: Num rows: 2214084911 Data size: 4626877672647 Basic stats: COMPLETE Column stats: NONE
                table:
                    input format: org.apache.hadoop.mapred.TextInputFormat
                    output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat
                    serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe

  Stage: Stage-0
    Fetch Operator
      limit: -1
      Processor Tree:
        ListSink

map: 两个TableScan

reduce: 两表join,然后再predicate: _col25 is null (type: boolean)

4. EXISTS examples

select * from (
    select stack(4, 
        'a', null, 'b', 'a'
    ) as key
) a
where not exists  (select key from (select stack(2, 'a', null) as key) b where a.key = b.key)

output:

NULL
b

5. LEFT JOIN query plan

select * from id_full as a
left join (select id from id_incr as b where day=20180512) as b
on a.id = b.id
where b.id is null
STAGE DEPENDENCIES:
  Stage-4 is a root stage , consists of Stage-1
  Stage-1
  Stage-0 depends on stages: Stage-1

STAGE PLANS:
  Stage: Stage-4
    Conditional Operator

  Stage: Stage-1
    Map Reduce
      Map Operator Tree:
          TableScan
            alias: b
            filterExpr: (day = 20180512) (type: boolean)
            Statistics: Num rows: 14933063 Data size: 15054692244 Basic stats: COMPLETE Column stats: NONE
            Select Operator
              expressions: id (type: string)
              outputColumnNames: _col0
              Statistics: Num rows: 14933063 Data size: 15054692244 Basic stats: COMPLETE Column stats: NONE
              Reduce Output Operator
                key expressions: _col0 (type: string)
                sort order: +
                Map-reduce partition columns: _col0 (type: string)
                Statistics: Num rows: 14933063 Data size: 15054692244 Basic stats: COMPLETE Column stats: NONE
          TableScan
            alias: a
            Statistics: Num rows: 4025608842 Data size: 8412504677023 Basic stats: COMPLETE Column stats: NONE
            Reduce Output Operator
              key expressions: id (type: string)
              sort order: +
              Map-reduce partition columns: id (type: string)
              Statistics: Num rows: 4025608842 Data size: 8412504677023 Basic stats: COMPLETE Column stats: NONE
              value expressions: id (type: string), ...
      Reduce Operator Tree:
        Join Operator
          condition map:
               Left Outer Join0 to 1
          keys:
            0 id (type: string)
            1 _col0 (type: string)
          outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9, _col10, _col11, _col12, _col13, _col14, _col15, _col16, _col17, _col18, _col19, _col24
          Statistics: Num rows: 4428169822 Data size: 9253755345295 Basic stats: COMPLETE Column stats: NONE
          Filter Operator
            predicate: _col24 is null (type: boolean)
            Statistics: Num rows: 2214084911 Data size: 4626877672647 Basic stats: COMPLETE Column stats: NONE
            Select Operator
              expressions: _col0 (type: string), _col1 (type: string), _col2 (type: string), _col3 (type: string), _col4 (type: string), _col5 (type: string), _col6 (type: string), _col7 (type: string), _col8 (type: string), _col9 (type: string), _col10 (type: string), _col11 (type: string), _col12 (type: string), _col13 (type: string), _col14 (type: string), _col15 (type: string), _col16 (type: string), _col17 (type: string), _col18 (type: string), _col29 (type: string), null (type: string)
              outputColumnNames: _col0, _col1, _col2, _col3, _col4, _col5, _col6, _col7, _col8, _col9, _col10, _col11, _col12, _col13, _col14, _col15, _col16, _col17, _col18, _col19, _col20
              Statistics: Num rows: 2214084911 Data size: 4626877672647 Basic stats: COMPLETE Column stats: NONE
              File Output Operator
                compressed: false
                Statistics: Num rows: 2214084911 Data size: 4626877672647 Basic stats: COMPLETE Column stats: NONE
                table:
                    input format: org.apache.hadoop.mapred.TextInputFormat
                    output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat
                    serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe

  Stage: Stage-0
    Fetch Operator
      limit: -1
      Processor Tree:
        ListSink

可以看到这个stage和NOT EXISTS的stage是一样的

6. 结论

在求差集上NOT EXISTSLEFT JOIN在执行上是等价的.

NOT IN需要分以下2种情况:

  1. 如果子查询的结果中没有NULL,则返回结果和NOT EXISTS,LEFT JOIN是一样的,只不过会多2个stage
  2. 如果子查询的结果中有NULL,则返回的结果为空,这点不同于NOT EXISTS,LEFT JOIN
    原文作者:丧诗
    原文地址: https://www.jianshu.com/p/406b9248b0cd
    本文转自网络文章,转载此文章仅为分享知识,如有侵权,请联系博主进行删除。
点赞