[SPARK-28275][SQL][PYTHON][TESTS] Convert and port 'count.sql' into UDF test base

[vinodkc]

What changes were proposed in this pull request?

This PR adds some tests converted from 'count.sql' to test UDFs

Diff comparing to 'count.sql'

diff --git a/sql/core/src/test/resources/sql-tests/results/count.sql.out b/sql/core/src/test/resources/sql-tests/results/udf/udf-count.sql.out
index b8a86d4c44..9476937abd 100644
--- a/sql/core/src/test/resources/sql-tests/results/count.sql.out
+++ b/sql/core/src/test/resources/sql-tests/results/udf/udf-count.sql.out
@@ -14,42 +14,42 @@ struct<>
 
 -- !query 1
 SELECT
-  count(*), count(1), count(null), count(a), count(b), count(a + b), count((a, b))
+  udf(count(*)), udf(count(1)), udf(count(null)), udf(count(a)), udf(count(b)), udf(count(a + b)), udf(count((a, b)))
 FROM testData
 -- !query 1 schema
-struct<count(1):bigint,count(1):bigint,count(NULL):bigint,count(a):bigint,count(b):bigint,count((a + b)):bigint,count(named_struct(a, a, b, b)):bigint>
+struct<udf(count(1)):string,udf(count(1)):string,udf(count(null)):string,udf(count(a)):string,udf(count(b)):string,udf(count((a + b))):string,udf(count(named_struct(a, a, b, b))):string>
 -- !query 1 output
 7	7	0	5	5	4	7
 
 
 -- !query 2
 SELECT
-  count(DISTINCT 1),
-  count(DISTINCT null),
-  count(DISTINCT a),
-  count(DISTINCT b),
-  count(DISTINCT (a + b)),
-  count(DISTINCT (a, b))
+  udf(count(DISTINCT 1)),
+  udf(count(DISTINCT null)),
+  udf(count(DISTINCT a)),
+  udf(count(DISTINCT b)),
+  udf(count(DISTINCT (a + b))),
+  udf(count(DISTINCT (a, b)))
 FROM testData
 -- !query 2 schema
-struct<count(DISTINCT 1):bigint,count(DISTINCT NULL):bigint,count(DISTINCT a):bigint,count(DISTINCT b):bigint,count(DISTINCT (a + b)):bigint,count(DISTINCT named_struct(a, a, b, b)):bigint>
+struct<udf(count(distinct 1)):string,udf(count(distinct null)):string,udf(count(distinct a)):string,udf(count(distinct b)):string,udf(count(distinct (a + b))):string,udf(count(distinct named_struct(a, a, b, b))):string>
 -- !query 2 output
 1	0	2	2	2	6
 
 
 -- !query 3
-SELECT count(a, b), count(b, a), count(testData.*) FROM testData
+SELECT udf(count(a, b)), udf(count(b, a)), udf(count(testData.*)) FROM testData
 -- !query 3 schema
-struct<count(a, b):bigint,count(b, a):bigint,count(a, b):bigint>
+struct<udf(count(a, b)):string,udf(count(b, a)):string,udf(count(a, b)):string>
 -- !query 3 output
 4	4	4
 
 
 -- !query 4
 SELECT
-  count(DISTINCT a, b), count(DISTINCT b, a), count(DISTINCT *), count(DISTINCT testData.*)
+  udf(count(DISTINCT a, b)), udf(count(DISTINCT b, a)), udf(count(DISTINCT *)), udf(count(DISTINCT testData.*))
 FROM testData
 -- !query 4 schema
-struct<count(DISTINCT a, b):bigint,count(DISTINCT b, a):bigint,count(DISTINCT a, b):bigint,count(DISTINCT a, b):bigint>
+struct<udf(count(distinct a, b)):string,udf(count(distinct b, a)):string,udf(count(distinct a, b)):string,udf(count(distinct a, b)):string>
 -- !query 4 output
 3	3	3	3

How was this patch tested?

Tested as guided in SPARK-27921.

[HyukjinKwon]

Hey @vinodkc, thanks for taking a look for this one. Actually we should add output files too :-). Mind double checking the guide I wrote at SPARK-27921 one by one?

[vinodkc]

@HyukjinKwon , Thanks for the guidance, I've added output file now.

[HyukjinKwon]

@vinodkc, can make some other conbinations like udf(count(*)), count(udf(a)), udf(count(udf(a)))?

[HyukjinKwon]

Here too :-)

[HyukjinKwon]

Looks fine if there are no output diff comparing to the original file

[SparkQA]

Test build #107427 has finished for PR 25089 at commit c7b4c46.

• This patch fails Spark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[SparkQA]

Test build #107429 has finished for PR 25089 at commit b173330.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[HyukjinKwon]

Merged to master.