spark-instrumented-optimizer/sql/hive-thriftserver
Wenchen Fan f719e9c200 [SPARK-36409][SQL][TESTS] Splitting test cases from datetime.sql
### What changes were proposed in this pull request?

Currently `datetime.sql` contains a lot of tests and will be run 3 times: default mode, ansi mode, ntz mode. It wastes the test time and also large test files are hard to read.

This PR proposes to split it into smaller ones:
1. `date.sql`, which contains date literals, functions and operations. It will be run twice with default and ansi mode.
2. `timestamp.sql`, which contains timestamp (no ltz or ntz suffix) literals, functions and operations. It will be run 4 times: default mode + ans off, defaul mode + ansi on, ntz mode + ansi off, ntz mode + ansi on.
3. `datetime_special.sql`, which create datetime values whose year is outside of [0, 9999]. This is a separated file as JDBC doesn't support them and need to ignore this test file. It will be run 4 times as well.
4. `timestamp_ltz.sql`, which contains timestamp_ltz literals and constructors. It will be run twice with default and ntz mode, to make sure its result doesn't change with the timestamp mode. Note that, operations with ltz are tested by `timestamp.sql`
5. `timestamp_ntz.sql`, which contains timestamp_ntz literals and constructors. It will be run twice with default and ntz mode, to make sure its result doesn't change with the timestamp mode. Note that, operations with ntz are tested by `timestamp.sql`

### Why are the changes needed?

reduce test run time.

### Does this PR introduce _any_ user-facing change?

no

### How was this patch tested?

N/A

Closes #33640 from cloud-fan/test.

Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Gengliang Wang <gengliang@apache.org>
2021-08-06 12:55:31 +08:00
..
src [SPARK-36409][SQL][TESTS] Splitting test cases from datetime.sql 2021-08-06 12:55:31 +08:00
pom.xml [SPARK-35838][BUILD][TESTS] Ensure all modules can be maven test independently in Scala 2.13 2021-06-22 06:31:24 -07:00