[SPARK-28271][SQL][PYTHON][TESTS] Convert and port 'pgSQL/aggregates_part2.sql' into UDF test base

[imback82]

What changes were proposed in this pull request?

This PR adds some tests converted from pgSQL/aggregates_part2.sql' to test UDFs. Please see contribution guide of this umbrella ticket - SPARK-27921.

Diff comparing to 'pgSQL/aggregates_part2.sql'

diff --git a/sql/core/src/test/resources/sql-tests/results/pgSQL/aggregates_part2.sql.out b/sql/core/src/test/resources/sql-tests/results/udf/pgSQL/udf-aggregates_part2.sql.out
index 2606d2eba7..00c06f94b5 100644
--- a/sql/core/src/test/resources/sql-tests/results/pgSQL/aggregates_part2.sql.out
+++ b/sql/core/src/test/resources/sql-tests/results/udf/pgSQL/udf-aggregates_part2.sql.out
@@ -57,23 +57,23 @@ true        false   true    false   true    true    true    true    true


 -- !query 3
-select min(unique1) from tenk1
+select min(udf(unique1)) from tenk1
 -- !query 3 schema
-struct<min(unique1):int>
+struct<min(udf(unique1)):string>
 -- !query 3 output
 0


 -- !query 4
-select max(unique1) from tenk1
+select udf(max(unique1)) from tenk1
 -- !query 4 schema
-struct<max(unique1):int>
+struct<udf(max(unique1)):string>
 -- !query 4 output
 9999


 -- !query 5
-select max(unique1) from tenk1 where unique1 < 42
+select max(unique1) from tenk1 where udf(unique1) < 42
 -- !query 5 schema
 struct<max(unique1):int>
 -- !query 5 output
@@ -81,7 +81,7 @@ struct<max(unique1):int>


 -- !query 6
-select max(unique1) from tenk1 where unique1 > 42
+select max(unique1) from tenk1 where unique1 > udf(42)
 -- !query 6 schema
 struct<max(unique1):int>
 -- !query 6 output
@@ -89,7 +89,7 @@ struct<max(unique1):int>


 -- !query 7
-select max(unique1) from tenk1 where unique1 > 42000
+select max(unique1) from tenk1 where udf(unique1) > 42000
 -- !query 7 schema
 struct<max(unique1):int>
 -- !query 7 output
@@ -97,7 +97,7 @@ NULL


 -- !query 8
-select max(tenthous) from tenk1 where thousand = 33
+select max(tenthous) from tenk1 where udf(thousand) = 33
 -- !query 8 schema
 struct<max(tenthous):int>
 -- !query 8 output
@@ -105,7 +105,7 @@ struct<max(tenthous):int>


 -- !query 9
-select min(tenthous) from tenk1 where thousand = 33
+select min(tenthous) from tenk1 where udf(thousand) = 33
 -- !query 9 schema
 struct<min(tenthous):int>
 -- !query 9 output
@@ -113,15 +113,15 @@ struct<min(tenthous):int>


 -- !query 10
-select distinct max(unique2) from tenk1
+select distinct max(udf(unique2)) from tenk1
 -- !query 10 schema
-struct<max(unique2):int>
+struct<max(udf(unique2)):string>
 -- !query 10 output
 9999


 -- !query 11
-select max(unique2) from tenk1 order by 1
+select max(unique2) from tenk1 order by udf(1)
 -- !query 11 schema
 struct<max(unique2):int>
 -- !query 11 output
@@ -129,7 +129,7 @@ struct<max(unique2):int>

 -- !query 12
-select max(unique2) from tenk1 order by max(unique2)
+select max(unique2) from tenk1 order by max(udf(unique2))
 -- !query 12 schema
 struct<max(unique2):int>
 -- !query 12 output
@@ -137,7 +137,7 @@ struct<max(unique2):int>


 -- !query 13
-select max(unique2) from tenk1 order by max(unique2)+1
+select udf(max(udf(unique2))) from tenk1 order by udf(max(unique2))+1
 -- !query 13 schema
-struct<max(unique2):int>
+struct<udf(max(udf(unique2))):string>
 -- !query 13 output
 9999


 -- !query 14
-select t1.max_unique2, g from (select max(unique2) as max_unique2 FROM tenk1) t1 LATERAL VIEW explode(array(1,2,3)) t2 AS g order by g desc
+select t1.max_unique2, udf(g) from (select max(udf(unique2)) as max_unique2 FROM tenk1) t1 LATERAL VIEW explode(array(1,2,3)) t2 AS g order by g desc
 -- !query 14 schema
-struct<max_unique2:int,g:int>
+struct<max_unique2:string,udf(g):string>
 -- !query 14 output
 9999   3
 9999   2
@@ -155,8 +155,8 @@ struct<max_unique2:int,g:int>


 -- !query 15
-select max(100) from tenk1
+select udf(max(100)) from tenk1
 -- !query 15 schema
-struct<max(100):int>
+struct<udf(max(100)):string>
 -- !query 15 output
 100

How was this patch tested?

Tested as guided in SPARK-27921.

[imback82]

cc: @HyukjinKwon

[dongjoon-hyun]

ok to test

[SparkQA]

Test build #107416 has finished for PR 25086 at commit 861f02a.

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

[HyukjinKwon]

Looks good otherwise.

[SparkQA]

Test build #107422 has finished for PR 25086 at commit 0550038.

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

[dongjoon-hyun]

Please use REL_12_BETA2 here because BETA2 is already out and more closer to REL_12. I guess REL_12_BETA2 will be the same with REL_12_BETA1.
If not, please make this up-to-date.

[HyukjinKwon]

Seems like there's one more case. Let me make a small PR separately to address this comment for aggregates_part2.sql as well.

[dongjoon-hyun]

+1.

[HyukjinKwon]

Merged to master.

Will make a follow right away, @dongjoon-hyun.

[dongjoon-hyun]

Thank you, @HyukjinKwon ~