Mailing-List: contact dev-help@drill.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@drill.apache.org
Date: Wed, 18 Jan 2017 16:58:26 +0000 (UTC)
From: "Paul Rogers (JIRA)" <jira@apache.org>
To: dev@drill.apache.org
Message-ID: <JIRA.13035927.1484758666000.42621.1484758706455@Atlassian.JIRA>
In-Reply-To: <JIRA.13035927.1484758666000@Atlassian.JIRA>
References: <JIRA.13035927.1484758666000@Atlassian.JIRA> <JIRA.13035927.1484758666562@jira-lw-us.apache.org>
Subject: [jira] [Created] (DRILL-5202) Planner misses opportunity to link
 sort & filter without remover
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
archived-at: Wed, 18 Jan 2017 16:58:31 -0000

Paul Rogers created DRILL-5202:
----------------------------------

             Summary: Planner misses opportunity to link sort & filter with=
out remover
                 Key: DRILL-5202
                 URL: https://issues.apache.org/jira/browse/DRILL-5202
             Project: Apache Drill
          Issue Type: Bug
    Affects Versions: 1.9.0
            Reporter: Paul Rogers
            Priority: Minor


Consider the following query:

{code}
SELECT * FROM (SELECT * FROM `mock`.`mock.json` ORDER BY col1) d WHERE d.co=
l1 =3D 'bogus'
{code}

The data source here is a mock: it simply generates a data set with 10 colu=
mns numbered col1 to col10. Then it generates 10,000 rows of data.

The plan for this query misses an optimization opportunity. (See plan below=
.)

The current plan is (abbreviated):

* scan
* ...
* sort
* selection vector remover
* project
* filter
* selection vector remover
* project
* screen

Careful inspection shows that this query is very simple. The following step=
s would work just as well:

* scan
* ...
* sort
* filter
* project
* screen

That is, the filter can handle an input with a selection vector. So, no SVR=
 is needed between the sort and the filter. Plus, this is a {{SELECT *}} qu=
ery, so all the extra projects don't really do anything useful, so they can=
 be removed where unneeded. The revised plan eliminates an unnecessary data=
 copy.

Of course, the planner should have pushed the filter below the sort. But th=
at is DRILL-5200.

{code}=20
 "graph" : [ {
    "pop" : "mock-scan",
    "@id" : 8,
    ...
  }, {
    "pop" : "project",
    "@id" : 7,
    "exprs" : [ {
      "ref" : "`T0=C2=A6=C2=A6*`",
      "expr" : "`*`"
    }, {
      "ref" : "`col1`",
      "expr" : "`col1`"
    } ],
    "child" : 8,
    ...
  }, {
    "pop" : "external-sort",
    "@id" : 6,
    "child" : 7,
    "orderings" : [ {
      "order" : "ASC",
      "expr" : "`col1`",
      "nullDirection" : "UNSPECIFIED"
    } ],
    ...
  }, {
    "pop" : "selection-vector-remover",
    "@id" : 5,
    "child" : 6,
    ...
  }, {
    "pop" : "project",
    "@id" : 4,
    "exprs" : [ {
      "ref" : "`T0=C2=A6=C2=A6*`",
      "expr" : "`T0=C2=A6=C2=A6*`"
    } ],
    "child" : 5,
    ...
  }, {
    "pop" : "filter",
    "@id" : 3,
    "child" : 4,
    ...
  }, {
    "pop" : "selection-vector-remover",
    "@id" : 2,
    ...
  }, {
    "pop" : "project",
    "@id" : 1,
    "exprs" : [ {
      "ref" : "`*`",
      "expr" : "`T0=C2=A6=C2=A6*`"
    } ],
    "child" : 2,
   ...
  }, {
    "pop" : "screen",
    "@id" : 0,
    ...
  } ]


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