hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chao Sun (JIRA)" <j...@apache.org>
Subject [jira] [Created] (HIVE-14630) Enable PPD for AND conditions when CBO is disabled
Date Thu, 25 Aug 2016 06:56:21 GMT
Chao Sun created HIVE-14630:
-------------------------------

             Summary: Enable PPD for AND conditions when CBO is disabled
                 Key: HIVE-14630
                 URL: https://issues.apache.org/jira/browse/HIVE-14630
             Project: Hive
          Issue Type: Bug
          Components: Logical Optimizer
    Affects Versions: 2.2.0
            Reporter: Chao Sun


Currently the PPD optimization seems not be able to handle AND conditions very well, when
CBO is not used. To illustrate with a example:

Table a:
|| col || type || part_col? ||
| id | int | no |
| datestr | string | yes |

Table b:
|| col || type || part_col? ||
| id | int | no |

And the following query:
{code}
SELECT a.id FROM a JOIN b
ON a.id = b.id
WHERE a.datestr >= '2016-08-20'
AND rand() > 0.5
{code}

For this query, the plan looks like the following:
{code}
STAGE DEPENDENCIES:
  Stage-1 is a root stage
  Stage-0 depends on stages: Stage-1

STAGE PLANS:
  Stage: Stage-1
    Map Reduce
      Map Operator Tree:
          TableScan
            alias: a
            Statistics: Num rows: 1 Data size: 0 Basic stats: PARTIAL Column stats: NONE
            Filter Operator
              predicate: id is not null (type: boolean)
              Statistics: Num rows: 1 Data size: 0 Basic stats: PARTIAL Column stats: NONE
              Reduce Output Operator
                key expressions: id (type: bigint)
                sort order: +
                Map-reduce partition columns: id (type: bigint)
                Statistics: Num rows: 1 Data size: 0 Basic stats: PARTIAL Column stats: NONE
                value expressions: datestr (type: string)
          TableScan
            alias: b
            Statistics: Num rows: 1 Data size: 0 Basic stats: PARTIAL Column stats: NONE
            Filter Operator
              predicate: id is not null (type: boolean)
              Statistics: Num rows: 1 Data size: 0 Basic stats: PARTIAL Column stats: NONE
              Reduce Output Operator
                key expressions: id (type: bigint)
                sort order: +
                Map-reduce partition columns: id (type: bigint)
                Statistics: Num rows: 1 Data size: 0 Basic stats: PARTIAL Column stats: NONE
      Reduce Operator Tree:
        Join Operator
          condition map:
               Inner Join 0 to 1
          keys:
            0 id (type: bigint)
            1 id (type: bigint)
          outputColumnNames: _col0, _col2
          Statistics: Num rows: 1 Data size: 0 Basic stats: PARTIAL Column stats: NONE
          Filter Operator
            predicate: ((_col2 >= '2016-08-20') and (rand() > 0.5)) (type: boolean)
            Statistics: Num rows: 1 Data size: 0 Basic stats: PARTIAL Column stats: NONE
            Select Operator
              expressions: _col0 (type: bigint)
              outputColumnNames: _col0
              Statistics: Num rows: 1 Data size: 0 Basic stats: PARTIAL Column stats: NONE
              File Output Operator
                compressed: false
                Statistics: Num rows: 1 Data size: 0 Basic stats: PARTIAL Column stats: NONE
                table:
                    input format: org.apache.hadoop.mapred.SequenceFileInputFormat
                    output format: org.apache.hadoop.hive.ql.io.HiveSequenceFileOutputFormat
                    serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe

  Stage: Stage-0
    Fetch Operator
      limit: -1
      Processor Tree:
        ListSink
{code}

Note that the predicate {{a.datestr >= '2016-08-20'}} is not pushed down, since {{rand()}}
is not deterministic and thus the whole predicate is not eligible.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message