[
https://issues.apache.org/jira/browse/CALCITE-5143?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17537818#comment-17537818
]
Julian Hyde commented on CALCITE-5143:
--------------------------------------
[~jiajunbernoulli], I agree that we shouldn't be making all of these time units
into keywords (reserved or otherwise). Could we change the implementation of
{{EXTRACT}} to look for an identifier as the first argument? "S" is a valid
identifier. "B" is a valid identifier, but {{EXTRACT(b FROM now())}} would be
an illegal call to {{EXTRACT}} because it is not a time unit supported by
Calcite.
It might be interesting to support time units such as {{{}MINUTE15{}}}, so that
{{FLOOR(x TO MINUTE15)}} would round down to 15 minutes (see [Looker's minutex
set of
types|https://docs.looker.com/reference/field-params/dimension_group#minutex]).
Moving from keywords to identifiers would make it easier to support such time
units.
> Planner#parse fails to parse a function when HOUR is passed as an argument
> --------------------------------------------------------------------------
>
> Key: CALCITE-5143
> URL: https://issues.apache.org/jira/browse/CALCITE-5143
> Project: Calcite
> Issue Type: Bug
> Components: babel
> Affects Versions: 1.30.0
> Reporter: Mitsunori Komatsu
> Assignee: Jiajun Xie
> Priority: Major
> Labels: pull-request-available
>
> I'm trying to parse and rewrite some SQL dialects including SQL Server which
> supports some functions that receive HOUR as an argument using. But Calcite's
> Planner#parse can't handle this kind of SQL.
> {code:java}
> FrameworkConfig config = Frameworks.newConfigBuilder().build();
> Planner planner = Frameworks.getPlanner(config);
> System.out.println(planner.parse("SELECT DATEDIFF(HOUR, NOW(), NOW())"));
> {code}
> {code:java}
> Exception in thread "main" org.apache.calcite.sql.parser.SqlParseException:
> Incorrect syntax near the keyword 'HOUR' at line 2, column 21.
> Was expecting one of:
> "ALL" ...
> "ARRAY" ...
> "CASE" ...
> :
> "GROUPING" ...
> "HOUR" ...
> "HOUR" "(" ...
> "*" ...
> ")" ...
>
> at
> org.apache.calcite.sql.parser.impl.SqlParserImpl.convertException(SqlParserImpl.java:389)
> at
> org.apache.calcite.sql.parser.impl.SqlParserImpl.normalizeException(SqlParserImpl.java:153)
> at
> org.apache.calcite.sql.parser.SqlParser.handleException(SqlParser.java:145)
> at org.apache.calcite.sql.parser.SqlParser.parseQuery(SqlParser.java:160)
> at org.apache.calcite.sql.parser.SqlParser.parseStmt(SqlParser.java:185)
> at org.apache.calcite.prepare.PlannerImpl.parse(PlannerImpl.java:214)
> at org.apache.calcite.tools.Planner.parse(Planner.java:50) {code}
> This issue happens with other date/time keyword-ish ones like YEAR.
> I guess it happens because HOUR and YEAR are reserved keywords.
> Is this an expected behavior? If so, is there any workaround?
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
