Piotr Bojko created CALCITE-2511:
------------------------------------
Summary: Possible bug in parsing implementation of rown_num select
Key: CALCITE-2511
URL: https://issues.apache.org/jira/browse/CALCITE-2511
Project: Calcite
Issue Type: Bug
Components: core
Affects Versions: 1.17.0, 1.18.0
Reporter: Piotr Bojko
Assignee: Julian Hyde
Attachments: implementation.java
I've found a possible bug.
With following view:
{code:json}
{
"name": "STEPS",
"type": "view",
"sql": [
"WITH ALLROWS AS ( ",
" SELECT ",
" ROW_NUMBER() OVER W AS IDX,
I.ID ISSUE, COALESCE(CG.AUTHOR, I.REPORTER) AS AUTHOR, COALESCE(CI.OLDVALUE,
I.STATUS) AS STATUS, I.CREATED AS STARTDATE, CG.CREATED AS ENDDATE ",
" FROM ISSUES AS I ",
" JOIN INTERNAL.CHANGEGROUP AS
CG ON CG.ISSUEID = I.ID ",
" JOIN INTERNAL.CHANGEITEM AS
CI ON CI.GROUPID = CG.ID ",
" WHERE CI.FIELDTYPE = 'jira'
AND CI.FIELD = 'status' ",
" WINDOW w AS (PARTITION BY
i.id ORDER BY CG.CREATED) ",
") ",
"SELECT ISSUE, AUTHOR, STATUS,
STARTDATE, ENDDATE ",
"FROM ALLROWS WHERE ALLROWS.IDX
<= 1 ",
"UNION ALL ",
"SELECT I.ID AS ISSUE,
CG.AUTHOR, CI.NEWVALUE AS STATUS, CG.CREATED AS STARTDATE, LEAD(CG.CREATED, 1)
OVER W AS ENDDATE ",
"FROM ISSUES i ",
"JOIN INTERNAL.CHANGEGROUP AS
CG ON CG.ISSUEID = I.ID ",
"JOIN INTERNAL.CHANGEITEM AS CI
ON CI.GROUPID = CG.ID ",
"WHERE CI.FIELDTYPE = 'jira'
AND CI.FIELD = 'status' ",
"WINDOW W AS (PARTITION BY I.ID
ORDER BY CG.CREATED) "
]
},
{code}
When performig
{code:sql}
SELECT * FROM STEPS
{code}
The result is
{noformat}
Line 1556, Column 0: Java$AmbiguousName is not allowed as an expression
statement. Expressions statements must be one of assignments, method
invocations, or object allocations..
{noformat}
As an attachment, I've attached implementation. Clearly the line 1555/1556 is
wrong...
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
