[
https://issues.apache.org/jira/browse/FLINK-28403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17562657#comment-17562657
]
suheng.cloud commented on FLINK-28403:
--------------------------------------
[~martijnvisser] thanks, actually I wonder whether it is a bug, so I think
maybe reported in Jira more suitable :)
cc [~jark] sorry to ping you in, can you help to have a look at this issue,
thanks~
> SqlParserPos value for SqlCall is ambiguous
> -------------------------------------------
>
> Key: FLINK-28403
> URL: https://issues.apache.org/jira/browse/FLINK-28403
> Project: Flink
> Issue Type: Bug
> Components: Table SQL / API
> Affects Versions: 1.15.0
> Reporter: suheng.cloud
> Priority: Minor
>
> Dear Community:
> For a long time I was puzzled by how SqlParserPos actually means in SqlCall,
> because I found the definitions for different statements in parserImpls.ftl
> are not unified.
> For example, in SqlCreateTable, we use `startPos.plus(getPos())` which means
> if I want to find out the end position of the last token, calling
> `SqlCreateTable::getParserPosition::getEndxx` is ok, but in some other
> statement like SqlRichDescribeTable, we just use the intermediate `pos` as
> the final position, so I must call
> `SqlRichDescribeTable.getParserPosition().plusAll(SqlRichDescribeTable::getOperandList())`
> to merge the result pos. Even worse in SqlStatementSet, which wrapped in
> SqlExecuate, both above methods not work because no child operand covered
> `END` syntax.
> My question is
> # Does the SqlParserPosition mean entire span for a sqlCall or not ?
> # What is the easyest and right way to get the "final position" of a
> SqlCall/SqlNode in fink sql?
> Correct me if I miss some point, thanks for any help.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
