Added SparkSQL support based on HiveDialect

[tanhaoqin]

SparkSQL is made to be almost compatible with HiveQL. However, the show tables syntax for Spark SQL is slightly different. As such, a new SparkSQL dialect is created based on the original HiveDialect with get_table_names modified to accommodate the results returned from Spark SQL's 'SHOW TABLES'.

[tanhaoqin]

This fixes https://github.com/dropbox/PyHive/issues/150 without breaking the original Hive engine.

[xqzhou]

Any progress on this?

[alexandrejuma]

+1

[kentore82]

+1 to get this PR merged :)

[maver1ck]

Do you thought also about changing get_view_names, get_table_names result based on isTemporary column ?

[tooptoop4]

bump

[mistercrunch]

@tanhaoqin this is failing the tests for some reason, maybe rebasing would fix the issue?

[tanhaoqin]

Changed some stuff so there's no longer any conflicts.

I am no longer working on the project that required this feature though. Didn't have time to write tests as I am not using any Hive/Presto on my side so the original Hive/Presto tests will fail anyway.

[gmcoringa]

Many tests done by sqlalchemy_test_case will fail due the lack of support of some types by spark (SPARK-21529).

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}