[
https://issues.apache.org/jira/browse/TAJO-757?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13967916#comment-13967916
]
Hyunsik Choi commented on TAJO-757:
-----------------------------------
Created a review request against branch master in reviewboard
https://reviews.apache.org/r/20289/
> Add parser rule and algebraic expression for window function
> ------------------------------------------------------------
>
> Key: TAJO-757
> URL: https://issues.apache.org/jira/browse/TAJO-757
> Project: Tajo
> Issue Type: Sub-task
> Components: distributed query plan, parser, physical operator,
> planner/optimizer
> Reporter: Hyunsik Choi
> Assignee: Hyunsik Choi
> Fix For: 1.0-incubating
>
> Attachments: TAJO-757.patch
>
>
> See the title. The object of this issue is to add ANTLR parser rule and
> algebraic expression for window function.
> According to SQL 2003 standards, BNF for window function is as follows:
> {noformat}
> <window function> ::= <window function type> OVER <window name or
> specification>
> <window function type> ::=
> <rank function type> <left paren> <right paren>
> | ROW_NUMBER <left paren> <right paren>
> | <aggregate function>
> <rank function type> ::= RANK | DENSE_RANK | PERCENT_RANK | CUME_DIST
> <window name or specification> ::= <window name> | <in-line window
> specification>
> <in-line window specification> ::= <window specification>
> <window specification> ::= <left paren> <window specification details>
> <right paren>
> <window specification details> ::=
> [ <existing window name> ] [ <window partition clause> ] [ <window
> order clause> ] [ <window frame clause> ]
> <existing window name> ::= <window name>
> <window partition clause> ::= PARTITION BY <window partition column
> reference list>
> <window partition column reference list> ::= <window partition column
> reference> [ { <comma> <window partition column reference> }... ]
> <window partition column reference> ::= <column reference> [ <collate
> clause> ]
> <window order clause> ::= ORDER BY <sort specification list>
> <window frame clause> ::= <window frame units> <window frame extent> [
> <window frame exclusion> ]
> <window frame units> ::= ROWS | RANGE
> <window frame extent> ::= <window frame start> | <window frame between>
> <window frame start> ::= UNBOUNDED PRECEDING | <window frame preceding>
> | CURRENT ROW
> <window frame preceding> ::= <unsigned value specification> PRECEDING
> <window frame between> ::= BETWEEN <window frame bound 1> AND <window
> frame bound 2>
> <window frame bound 1> ::= <window frame bound>
> <window frame bound 2> ::= <window frame bound>
> <window frame bound> ::=
> <window frame start>
> | UNBOUNDED FOLLOWING
> | <window frame following>
> <window frame following> ::= <unsigned value specification> FOLLOWING
> <window frame exclusion> ::=
> EXCLUDE CURRENT ROW
> | EXCLUDE GROUP
> | EXCLUDE TIES
> | EXCLUDE NO OTHERS
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.2#6252)
