Steven Talbot created CALCITE-5147: -------------------------------------- Summary: Parser configured with BigQuery dialect cannot parse timestamp literal Key: CALCITE-5147 URL: https://issues.apache.org/jira/browse/CALCITE-5147 Project: Calcite Issue Type: Bug Reporter: Steven Talbot
It looks like [https://github.com/apache/calcite/commit/c1052b343724c4a95bcf25419b70bc6032e0846b#diff-e873041549333502af52ece8a1b34301ae5a059ff4719e9bddbaef48929e7047R7796] creates a special string literal token for BigQuery. But that means that (AFAICT) [https://github.com/apache/calcite/blob/b9c2099ea92a575084b55a206efc5dd341c0df62/core/src/main/codegen/templates/Parser.jj#L4529] then does not find the expected token type after a TIMESTAMP to make a timestamp literal, so something as simple as {code:java} SqlParser.create("SELECT TIMESTAMP '2018-02-17 13:22:04'", BigQuerySqlDialect.DEFAULT.configureParser(SqlParser.config())).parseQuery() {code} fails. Probably most other places in Parser.jj that use the "<QUOTED_STRING>" token suffer from the same issue. -- This message was sent by Atlassian Jira (v8.20.7#820007)