dawidwys commented on code in PR #23633: URL: https://github.com/apache/flink/pull/23633#discussion_r1380400786
########## flink-table/flink-table-planner/src/main/scala/org/apache/flink/table/planner/typeutils/TypeCoercion.scala: ########## @@ -53,85 +53,4 @@ object TypeCoercion { case _ => false } - - /** - * All the supported cast types in flink-table. - * - * Note: No distinction between explicit and implicit conversions Note: This is a subset of - * SqlTypeAssignmentRule Note: This may lose type during the cast. - */ - def canCast(from: LogicalType, to: LogicalType): Boolean = - (from.getTypeRoot, to.getTypeRoot) match { - case (_, _) if from == to => true - - case (_, VARCHAR | CHAR) => true - - case (VARCHAR | CHAR, _) if isNumeric(to) => true - case (VARCHAR | CHAR, BOOLEAN) => true - case (VARCHAR | CHAR, DECIMAL) => true - case (VARCHAR | CHAR, DATE) => true - case (VARCHAR | CHAR, TIME_WITHOUT_TIME_ZONE) => true - case (VARCHAR | CHAR, TIMESTAMP_WITHOUT_TIME_ZONE) => true - case (VARCHAR | CHAR, TIMESTAMP_WITH_LOCAL_TIME_ZONE) => true - - case (BOOLEAN, _) if isNumeric(to) => true - case (BOOLEAN, DECIMAL) => true - case (_, BOOLEAN) if isNumeric(from) => true - case (DECIMAL, BOOLEAN) => true - - case (_, _) if isNumeric(from) && isNumeric(to) => true - case (_, DECIMAL) if isNumeric(from) => true - case (DECIMAL, _) if isNumeric(to) => true - case (DECIMAL, DECIMAL) => true - case (INTEGER, DATE) => true - case (INTEGER, TIME_WITHOUT_TIME_ZONE) => true - case (TINYINT, TIMESTAMP_WITHOUT_TIME_ZONE) => true - case (SMALLINT, TIMESTAMP_WITHOUT_TIME_ZONE) => true - case (INTEGER, TIMESTAMP_WITHOUT_TIME_ZONE) => true - case (BIGINT, TIMESTAMP_WITHOUT_TIME_ZONE) => true - case (DOUBLE, TIMESTAMP_WITHOUT_TIME_ZONE) => true - case (FLOAT, TIMESTAMP_WITHOUT_TIME_ZONE) => true - case (INTEGER, INTERVAL_YEAR_MONTH) => true - case (BIGINT, INTERVAL_DAY_TIME) => true - - case (DATE, TIME_WITHOUT_TIME_ZONE) => false - case (TIME_WITHOUT_TIME_ZONE, DATE) => false - case (_, _) if isTimePoint(from) && isTimePoint(to) => true - case (DATE, INTEGER) => true - case (TIME_WITHOUT_TIME_ZONE, INTEGER) => true - case (TIMESTAMP_WITHOUT_TIME_ZONE, TINYINT) => true - case (TIMESTAMP_WITHOUT_TIME_ZONE, INTEGER) => true - case (TIMESTAMP_WITHOUT_TIME_ZONE, SMALLINT) => true - case (TIMESTAMP_WITHOUT_TIME_ZONE, BIGINT) => true - case (TIMESTAMP_WITHOUT_TIME_ZONE, DOUBLE) => true - case (TIMESTAMP_WITHOUT_TIME_ZONE, FLOAT) => true - - case (INTERVAL_YEAR_MONTH, INTEGER) => true - case (INTERVAL_DAY_TIME, BIGINT) => true - - case _ => false - } - - /** All the supported reinterpret types in flink-table. */ - def canReinterpret(from: LogicalType, to: LogicalType): Boolean = - (from.getTypeRoot, to.getTypeRoot) match { - case (_, _) if from == to => true - - case (DATE, INTEGER) => true - case (TIME_WITHOUT_TIME_ZONE, INTEGER) => true - case (TIMESTAMP_WITHOUT_TIME_ZONE, BIGINT) => true - case (INTEGER, DATE) => true - case (INTEGER, TIME_WITHOUT_TIME_ZONE) => true - case (BIGINT, TIMESTAMP_WITHOUT_TIME_ZONE) => true - case (INTEGER, INTERVAL_YEAR_MONTH) => true - case (BIGINT, INTERVAL_DAY_TIME) => true - case (INTERVAL_YEAR_MONTH, INTEGER) => true - case (INTERVAL_DAY_TIME, BIGINT) => true - - case (DATE, BIGINT) => true - case (TIME_WITHOUT_TIME_ZONE, BIGINT) => true - case (INTERVAL_YEAR_MONTH, BIGINT) => true - - case _ => false Review Comment: `canCast` method I removed because it was not used Now, answering you question. The way I understand it. `REINTERPRET_CAST` is an internal function, that requires knowledge of the internal data formats. It says to change the logical type without touching the underlying data format. It can be used if both logical types use the same physical format. `CAST` can actually modify data. E.g. you can cast an INT to STRING which changes the underlying format etc. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@flink.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org