问题:

Do we know any syntax reference for spark-sql? I mean the style like MySQL SELECT syntax, or PostgreSQL Documentation: SELECT.

If not, what is the best source we can access?

回答1:

SparkSql follows Hive style, so you can refer to Hive Syntax for better documentation.

The supported and unsupported Hive features by SparkSql can be found in the official documentation.

回答2:

If not, what is the best source we can access?

I don't know about any better source that could be considered "the best" but the source code of Apache Spark itself. I can imagine that is not as user-friendly as the mentioned sources, but Apache Spark is a highly dynamic open source project and it'd be very hard to keep up with all the changes.

If you're asking about what's supported in so-called SQL mode, you should consult SqlBase.g4 which is the ANTLR grammar of Spark SQL. Unless there's a bug in ANTLR you cannot do anything but what's in the grammar.