[
https://issues.apache.org/jira/browse/CALCITE-3147?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16878236#comment-16878236
]
Julian Hyde commented on CALCITE-3147:
--------------------------------------
After these changes, BigQuery dialect:
* quotes identifiers only if it needs to (i.e. if they contain characters that
are not A-Z, a-z, 0-9 or _; or are reserved keywords);
* quotes character literals using back-slash for embedded single-quotes;
* generates OFFSET and LIMIT rather than FETCH;
* generates "ORDER BY x IS NULL, x" rather than "ORDER BY x NULLS LAST",
because BigQuery does not support NULLS FIRST or NULLS LAST.
> BigQuery (standard SQL) quoting and identifier issues
> -----------------------------------------------------
>
> Key: CALCITE-3147
> URL: https://issues.apache.org/jira/browse/CALCITE-3147
> Project: Calcite
> Issue Type: Bug
> Reporter: Steven Talbot
> Priority: Major
>
> BigQuery escapes single quotes with backticks, and requires triple-quote
> quoting (like Python) when there is a newline. Also, it does not support the
> same set of characters in identifiers as other dialects of SQL. For example,
> it does not support '$'. Ideally, RelToSqlConverter (with the logic of the
> BigQuery dialect) would handle all this.
>
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
