[GitHub] [spark] cloud-fan commented on issue #27107: [SPARK-30436][SQL] Allow CREATE EXTERNAL TABLE with only requiring LOCATION

cloud-fan commented on issue #27107: [SPARK-30436][SQL] Allow CREATE EXTERNAL TABLE with only requiring LOCATION URL: https://github.com/apache/spark/pull/27107#issuecomment-571491435 You are right that these two parser rules are order sensitive. It's better to make `createHiveTable`

[GitHub] [spark] cloud-fan commented on issue #27107: [SPARK-30436][SQL] Allow CREATE EXTERNAL TABLE with only requiring LOCATION

cloud-fan commented on issue #27107: [SPARK-30436][SQL] Allow CREATE EXTERNAL TABLE with only requiring LOCATION URL: https://github.com/apache/spark/pull/27107#issuecomment-571454410 > And then I don't see a reason to retain EXTERNAL to first syntax as it will just make statement fail.

[GitHub] [spark] cloud-fan commented on issue #27107: [SPARK-30436][SQL] Allow CREATE EXTERNAL TABLE with only requiring LOCATION

cloud-fan commented on issue #27107: [SPARK-30436][SQL] Allow CREATE EXTERNAL TABLE with only requiring LOCATION URL: https://github.com/apache/spark/pull/27107#issuecomment-571443983 There are 2 kinds of syntaxes: `visitCreateTable` and `visitCreateHiveTable`. The major difference

[GitHub] [spark] cloud-fan commented on issue #27107: [SPARK-30436][SQL] Allow CREATE EXTERNAL TABLE with only requiring LOCATION

cloud-fan commented on issue #27107: [SPARK-30436][SQL] Allow CREATE EXTERNAL TABLE with only requiring LOCATION URL: https://github.com/apache/spark/pull/27107#issuecomment-571430471 I think the question here is what shall we do for `CREATE EXTERNAL TABLE ... LOCATION`. In Spark 2.4 it