spark-instrumented-optimizer/sql/catalyst
hyukjinkwon 33e337c118 [SPARK-24709][SQL][FOLLOW-UP] Make schema_of_json's input json as literal only
## What changes were proposed in this pull request?

The main purpose of `schema_of_json` is the usage of combination with `from_json` (to make up the leak of schema inference) which takes its schema only as literal; however, currently `schema_of_json` allows JSON input as non-literal expressions (e.g, column).

This was mistakenly allowed - we don't have to take other usages rather then the main purpose into account for now.

This PR makes a followup to only allow literals for `schema_of_json`'s JSON input. We can allow non literal expressions later when it's needed or there are some usecase for it.

## How was this patch tested?

Unit tests were added.

Closes #22775 from HyukjinKwon/SPARK-25447-followup.

Lead-authored-by: hyukjinkwon <gurwls223@apache.org>
Co-authored-by: Hyukjin Kwon <gurwls223@apache.org>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
2018-10-26 22:14:43 +08:00
..
benchmarks [SPARK-25657][SQL][TEST] Refactor HashBenchmark to use main method 2018-10-07 09:49:37 -07:00
src [SPARK-24709][SQL][FOLLOW-UP] Make schema_of_json's input json as literal only 2018-10-26 22:14:43 +08:00
pom.xml [SPARK-25393][SQL] Adding new function from_csv() 2018-10-17 09:32:05 +08:00