From 6328ef2aac3201e933c182007c4644677cbc43a4 Mon Sep 17 00:00:00 2001
From: Michael Xu <xzhseh@gmail.com>
Date: Thu, 1 Feb 2024 20:59:06 -0500
Subject: [PATCH] refactor sql udf test

---
 e2e_test/udf/sql_udf.slt | 428 ++++++++++++++++++++-------------------
 1 file changed, 225 insertions(+), 203 deletions(-)

diff --git a/e2e_test/udf/sql_udf.slt b/e2e_test/udf/sql_udf.slt
index 758ec43ca53fc..f2db14324b0eb 100644
--- a/e2e_test/udf/sql_udf.slt
+++ b/e2e_test/udf/sql_udf.slt
@@ -1,325 +1,348 @@
 statement ok
 SET RW_IMPLICIT_FLUSH TO true;
 
+#############################################
+# Basic tests for anonymous / named sql udf #
+#############################################
+
 # Create an anonymous function with double dollar as clause
 statement ok
 create function add(INT, INT) returns int language sql as $$select $1 + $2$$;
 
+query I
+select add(1, -1);
+----
+0
+
 # Create an anonymous function with single quote as clause
 statement ok
 create function sub(INT, INT) returns int language sql as 'select $1 - $2';
 
+query I
+select sub(1, 1);
+----
+0
+
+
+
 # Create an anonymous function that calls other pre-defined sql udfs
 statement ok
 create function add_sub_binding() returns int language sql as 'select add(1, 1) + sub(2, 2)';
 
-# Create a named sql udf
+query I
+select add_sub_binding();
+----
+2
+
+# Use them all together
+query III
+select add(1, -1), sub(1, 1), add_sub_binding();
+----
+0 0 2
+
+# Create a named sql udf with single quote as clause
 statement ok
 create function add_named(a INT, b INT) returns int language sql as 'select a + b';
 
-# Create another named sql udf
-statement ok
-create function sub_named(a INT, b INT) returns int language sql as 'select a - b';
+query I
+select add_named(1, -1);
+----
+0
 
-# Mixed parameter with named / anonymous parameters
+# Create another named sql udf with double dollar as clause
 statement ok
-create function add_sub_mix(INT, a INT, INT) returns int language sql as 'select $1 - a + $3';
+create function sub_named(a INT, b INT) returns int language sql as $$select a - b$$;
 
-# Mixed parameter with calling inner sql udfs
-# statement ok
-# create function add_sub_mix_wrapper(INT, a INT, INT) returns int language sql as 'select add($1, a) + a + sub(a, $3)';
+query I
+select sub_named(1, 1);
+----
+0
 
-# Named sql udf with corner case
+# Mixed parameter with named / anonymous parameters
 statement ok
-create function corner_case(INT, a INT, INT) returns varchar language sql as $$select '$1 + a + $3'$$;
+create function add_sub_mix(INT, a INT, INT) returns int language sql as 'select $1 - a + $3';
 
-# Named sql udf with invalid parameter in body definition
-# Will be rejected at creation time
-statement error failed to find named parameter aa
-create function unknown_parameter(a INT) returns int language sql as 'select a + aa + a';
+query I
+select add_sub_mix(1, 2, 3);
+----
+2
 
 # Call anonymous sql udf inside named sql udf
 statement ok
 create function add_named_wrapper(a INT, b INT) returns int language sql as 'select add(a, b)';
 
-# Create an anonymous function that calls built-in functions
-# Note that double dollar signs should be used otherwise the parsing will fail, as illutrates below
-statement ok
-create function call_regexp_replace() returns varchar language sql as $$select regexp_replace('💩💩💩💩💩foo🤔️bar亲爱的😭baz这不是爱情❤️‍🔥', 'baz(...)', '这是🥵', 'ic')$$;
-
-statement error Expected end of statement, found: 💩
-create function call_regexp_replace() returns varchar language sql as 'select regexp_replace('💩💩💩💩💩foo🤔️bar亲爱的😭baz这不是爱情❤️‍🔥', 'baz(...)', '这是🥵', 'ic')';
+query I
+select add_named_wrapper(1, -1);
+----
+0
 
 # Create an anonymous function with return expression
 statement ok
 create function add_return(INT, INT) returns int language sql return $1 + $2;
 
+query I
+select add_return(1, 1);
+----
+2
+
 statement ok
 create function add_return_binding() returns int language sql return add_return(1, 1) + add_return(1, 1);
 
-# Recursive definition can NOT be accepted at present due to semantic check
-statement error failed to conduct semantic check, please see if you are calling non-existence functions
-create function recursive(INT, INT) returns int language sql as 'select recursive($1, $2) + recursive($1, $2)';
-
-# Complex but error-prone definition, recursive & normal sql udfs interleaving
-statement error failed to conduct semantic check, please see if you are calling non-existence functions
-create function recursive_non_recursive(INT, INT) returns int language sql as 'select recursive($1, $2) + sub($1, $2)';
+query I
+select add_return_binding();
+----
+4
 
-# Recursive corner case
 statement ok
-create function foo(INT) returns varchar language sql as $$select 'foo(INT)'$$;
+create function print(INT) returns int language sql as 'select $1';
 
-# Create a wrapper function for `add` & `sub`
+query T
+select print(114514);
+----
+114514
+
+# Multiple type interleaving sql udf
 statement ok
-create function add_sub_wrapper(INT, INT) returns int language sql as 'select add($1, $2) + sub($1, $2) + 114512';
+create function add_sub(INT, FLOAT, INT) returns float language sql as $$select -$1 + $2 - $3$$;
 
-# Create a valid recursive function
-# Please note we do NOT support actual running the recursive sql udf at present
-statement error failed to conduct semantic check, please see if you are calling non-existence functions
-create function fib(INT) returns int
- language sql as 'select case
-        when $1 = 0 then 0
-        when $1 = 1 then 1
-        when $1 = 2 then 1
-        when $1 = 3 then 2
-      else fib($1 - 1) + fib($1 - 2)
-      end;';
+query I
+select add_sub(1, 5.1415926, 1);
+----
+3.1415926
 
-# The execution will eventually exceed the pre-defined max stack depth
-# statement error function fib calling stack depth limit exceeded
-# select fib(100);
+query III
+select add(1, -1), sub(1, 1), add_sub(1, 5.1415926, 1);
+----
+0 0 3.1415926
 
-# Currently create a materialized view with a recursive sql udf will be rejected
-# statement error function fib calling stack depth limit exceeded
-# create materialized view foo_mv as select fib(100);
+# Complex types interleaving
+statement ok
+create function add_sub_types(INT, BIGINT, FLOAT, DECIMAL, REAL) returns double language sql as 'select $1 + $2 - $3 + $4 + $5';
+
+query I
+select add_sub_types(1, 1919810114514, 3.1415926, 1.123123, 101010.191919);
+----
+1919810215523.1734494
 
 statement ok
-create function regexp_replace_wrapper(varchar) returns varchar language sql as $$select regexp_replace($1, 'baz(...)', '这是🥵', 'ic')$$;
+create function add_sub_return(INT, FLOAT, INT) returns float language sql return -$1 + $2 - $3;
 
+query I
+select add_sub_return(1, 5.1415926, 1);
+----
+3.1415926
+
+# Create a wrapper function for `add` & `sub`
 statement ok
-create function print(INT) returns int language sql as 'select $1';
+create function add_sub_wrapper(INT, INT) returns int language sql as 'select add($1, $2) + sub($1, $2) + 114512';
+
+query I
+select add_sub_wrapper(1, 1);
+----
+114514
+
+##########################################################
+# Basic sql udfs integrated with the use of mock tables  #
+# P.S. This is also a simulation of real world use cases #
+##########################################################
 
-# Adjust the input value of the calling function (i.e., `print` here) with the actual input parameter
 statement ok
-create function print_add_one(INT) returns int language sql as 'select print($1 + 1)';
+create table t1 (c1 INT, c2 INT);
 
 statement ok
-create function print_add_two(INT) returns int language sql as 'select print($1 + $1)';
+create table t2 (c1 INT, c2 FLOAT, c3 INT);
 
-# Calling a non-existence function
-statement error failed to conduct semantic check, please see if you are calling non-existence functions
-create function non_exist(INT) returns int language sql as 'select yo(114514)';
+# Special table for named sql udf
+statement ok
+create table t3 (a INT, b INT);
 
-# Try to create an anonymous sql udf whose return type mismatches with the sql body definition
-statement error return type mismatch detected
-create function type_mismatch(INT) returns varchar language sql as 'select $1 + 114514 + $1';
+statement ok
+insert into t1 values (1, 1), (2, 2), (3, 3), (4, 4), (5, 5);
 
-# A valid example
 statement ok
-create function type_match(INT) returns varchar language sql as $$select '$1 + 114514 + $1'$$;
+insert into t2 values (1, 3.14, 2), (2, 4.44, 5), (20, 10.30, 02);
 
-query T
-select type_match(114514);
-----
-$1 + 114514 + $1
+statement ok
+insert into t3 values (1, 1), (2, 2), (3, 3), (4, 4), (5, 5);
 
-# Call the defined anonymous sql udfs
 query I
-select add(1, -1);
+select c1, c2, add_return(c1, c2) from t1 order by c1 asc;
 ----
-0
+1 1 2
+2 2 4
+3 3 6
+4 4 8
+5 5 10
 
-query I
-select sub(1, 1);
+query III
+select sub(c1, c2), c1, c2, add(c1, c2) from t1 order by c1 asc;
 ----
-0
+0 1 1 2
+0 2 2 4
+0 3 3 6
+0 4 4 8
+0 5 5 10
 
-# Call the defined named sql udfs
-query I
-select add_named(1, -1);
+query IIIIII
+select c1, c2, c3, add(c1, c3), sub(c1, c3), add_sub(c1, c2, c3) from t2 order by c1 asc;
 ----
-0
+1 3.14 2 3 -1 0.14000000000000012
+2 4.44 5 7 -3 -2.5599999999999996
+20 10.3 2 22 18 -11.7
 
-query I
-select sub_named(1, 1);
+query IIIIII
+select c1, c2, c3, add(c1, c3), sub(c1, c3), add_sub_return(c1, c2, c3) from t2 order by c1 asc;
 ----
-0
+1 3.14 2 3 -1 0.14000000000000012
+2 4.44 5 7 -3 -2.5599999999999996
+20 10.3 2 22 18 -11.7
 
 query I
-select add_sub_mix(1, 2, 3);
+select add_named(a, b) from t3 order by a asc;
 ----
 2
+4
+6
+8
+10
 
-query T
-select corner_case(1, 2, 3);
-----
-$1 + a + $3
-
-query I
-select add_named_wrapper(1, -1);
-----
-0
+##############################
+# Corner & Special cases tests #
+##############################
 
-query I
-select add_sub_binding();
-----
-2
+# Mixed parameter with calling inner sql udfs
+# statement ok
+# create function add_sub_mix_wrapper(INT, a INT, INT) returns int language sql as 'select add($1, a) + a + sub(a, $3)';
 
-query III
-select add(1, -1), sub(1, 1), add_sub_binding();
-----
-0 0 2
+# Named sql udf with corner case
+statement ok
+create function corner_case(INT, a INT, INT) returns varchar language sql as $$select '$1 + a + $3'$$;
 
-query I
-select add_return(1, 1);
+query T
+select corner_case(1, 2, 3);
 ----
-2
+$1 + a + $3
 
-query I
-select add_return_binding();
-----
-4
+# Create an anonymous function that calls built-in functions
+# Note that double dollar signs should be used otherwise the parsing will fail, as illutrates below
+statement ok
+create function call_regexp_replace() returns varchar language sql as $$select regexp_replace('💩💩💩💩💩foo🤔️bar亲爱的😭baz这不是爱情❤️‍🔥', 'baz(...)', '这是🥵', 'ic')$$;
 
 query T
 select call_regexp_replace();
 ----
 💩💩💩💩💩foo🤔️bar亲爱的😭这是🥵爱情❤️‍🔥
 
+statement ok
+create function regexp_replace_wrapper(varchar) returns varchar language sql as $$select regexp_replace($1, 'baz(...)', '这是🥵', 'ic')$$;
+
 query T
 select regexp_replace_wrapper('💩💩💩💩💩foo🤔️bar亲爱的😭baz这不是爱情❤️‍🔥');
 ----
 💩💩💩💩💩foo🤔️bar亲爱的😭这是🥵爱情❤️‍🔥
 
+# Recursive corner case (i.e., valid definition should not be rejected)
+statement ok
+create function foo(INT) returns varchar language sql as $$select 'foo(INT)'$$;
+
 query T
 select foo(114514);
 ----
 foo(INT)
 
-# Rejected deep calling stack
-# statement error function recursive calling stack depth limit exceeded
-# select recursive(1, 1);
-
-# Same as above
-# statement error function recursive calling stack depth limit exceeded
-# select recursive_non_recursive(1, 1);
+# Adjust the input value of the calling function (i.e., `print` here) with the actual input parameter
+statement ok
+create function print_add_one(INT) returns int language sql as 'select print($1 + 1)';
 
-query I
-select add_sub_wrapper(1, 1);
-----
-114514
+statement ok
+create function print_add_two(INT) returns int language sql as 'select print($1 + $1)';
 
 query III
 select print_add_one(1), print_add_one(114513), print_add_two(2);
 ----
 2 114514 4
 
-# Create a mock table for anonymous sql udf
+# Note: the valid example of `type_mismatch` in the below test section
 statement ok
-create table t1 (c1 INT, c2 INT);
+create function type_match(INT) returns varchar language sql as $$select '$1 + 114514 + $1'$$;
 
-# Create a mock table for named sql udf
-statement ok
-create table t3 (a INT, b INT);
+query T
+select type_match(114514);
+----
+$1 + 114514 + $1
 
-# Insert some data into the mock table
-statement ok
-insert into t1 values (1, 1), (2, 2), (3, 3), (4, 4), (5, 5);
+#################################################################################
+# Invalid definition (and maybe not yet supported features 🤪) / use case tests #
+#################################################################################
 
-statement ok
-insert into t3 values (1, 1), (2, 2), (3, 3), (4, 4), (5, 5);
+# Named sql udf with invalid parameter in body definition
+# Will be rejected at creation time
+statement error failed to find named parameter aa
+create function unknown_parameter(a INT) returns int language sql as 'select a + aa + a';
 
-query I
-select add_named(a, b) from t3 order by a asc;
-----
-2
-4
-6
-8
-10
+statement error Expected end of statement, found: 💩
+create function call_regexp_replace() returns varchar language sql as 'select regexp_replace('💩💩💩💩💩foo🤔️bar亲爱的😭baz这不是爱情❤️‍🔥', 'baz(...)', '这是🥵', 'ic')';
 
-query III
-select sub(c1, c2), c1, c2, add(c1, c2) from t1 order by c1 asc;
-----
-0 1 1 2
-0 2 2 4
-0 3 3 6
-0 4 4 8
-0 5 5 10
+# Recursive definition can NOT be accepted at present due to semantic check
+statement error failed to conduct semantic check, please see if you are calling non-existence functions
+create function recursive(INT, INT) returns int language sql as 'select recursive($1, $2) + recursive($1, $2)';
 
-query I
-select c1, c2, add_return(c1, c2) from t1 order by c1 asc;
-----
-1 1 2
-2 2 4
-3 3 6
-4 4 8
-5 5 10
+# Complex but error-prone definition, recursive & normal sql udfs interleaving
+statement error failed to conduct semantic check, please see if you are calling non-existence functions
+create function recursive_non_recursive(INT, INT) returns int language sql as 'select recursive($1, $2) + sub($1, $2)';
 
-# Recursive sql udf with normal table
+# Create a valid recursive function
+# Please note we do NOT support actual running the recursive sql udf at present
+statement error failed to conduct semantic check, please see if you are calling non-existence functions
+create function fib(INT) returns int
+ language sql as 'select case
+        when $1 = 0 then 0
+        when $1 = 1 then 1
+        when $1 = 2 then 1
+        when $1 = 3 then 2
+      else fib($1 - 1) + fib($1 - 2)
+      end;';
+
+# The execution will eventually exceed the pre-defined max stack depth
 # statement error function fib calling stack depth limit exceeded
-# select fib(c1) from t1;
+# select fib(100);
 
-# Recursive sql udf with materialized view
+# Currently create a materialized view with a recursive sql udf will be rejected
 # statement error function fib calling stack depth limit exceeded
-# create materialized view bar_mv as select fib(c1) from t1;
+# create materialized view foo_mv as select fib(100);
+
+# Calling a non-existence function
+statement error failed to conduct semantic check, please see if you are calling non-existence functions
+create function non_exist(INT) returns int language sql as 'select yo(114514)';
+
+# Try to create an anonymous sql udf whose return type mismatches with the sql body definition
+statement error return type mismatch detected
+create function type_mismatch(INT) returns varchar language sql as 'select $1 + 114514 + $1';
 
 # Invalid function body syntax
 statement error Expected an expression:, found: EOF at the end
 create function add_error(INT, INT) returns int language sql as $$select $1 + $2 +$$;
 
-# Multiple type interleaving sql udf
-statement ok
-create function add_sub(INT, FLOAT, INT) returns float language sql as $$select -$1 + $2 - $3$$;
-
-# Complex types interleaving
-statement ok
-create function add_sub_types(INT, BIGINT, FLOAT, DECIMAL, REAL) returns double language sql as 'select $1 + $2 - $3 + $4 + $5';
-
-statement ok
-create function add_sub_return(INT, FLOAT, INT) returns float language sql return -$1 + $2 - $3;
-
-query I
-select add_sub(1, 5.1415926, 1);
-----
-3.1415926
-
-query I
-select add_sub_return(1, 5.1415926, 1);
-----
-3.1415926
-
-query III
-select add(1, -1), sub(1, 1), add_sub(1, 5.1415926, 1);
-----
-0 0 3.1415926
-
-query I
-select add_sub_types(1, 1919810114514, 3.1415926, 1.123123, 101010.191919);
-----
-1919810215523.1734494
+# Rejected deep calling stack
+# statement error function recursive calling stack depth limit exceeded
+# select recursive(1, 1);
 
-# Create another mock table
-statement ok
-create table t2 (c1 INT, c2 FLOAT, c3 INT);
+# Same as above
+# statement error function recursive calling stack depth limit exceeded
+# select recursive_non_recursive(1, 1);
 
-statement ok
-insert into t2 values (1, 3.14, 2), (2, 4.44, 5), (20, 10.30, 02);
+# Recursive sql udf with normal table
+# statement error function fib calling stack depth limit exceeded
+# select fib(c1) from t1;
 
-query IIIIII
-select c1, c2, c3, add(c1, c3), sub(c1, c3), add_sub(c1, c2, c3) from t2 order by c1 asc;
-----
-1 3.14 2 3 -1 0.14000000000000012
-2 4.44 5 7 -3 -2.5599999999999996
-20 10.3 2 22 18 -11.7
+# Recursive sql udf with materialized view
+# statement error function fib calling stack depth limit exceeded
+# create materialized view bar_mv as select fib(c1) from t1;
 
-query IIIIII
-select c1, c2, c3, add(c1, c3), sub(c1, c3), add_sub_return(c1, c2, c3) from t2 order by c1 asc;
-----
-1 3.14 2 3 -1 0.14000000000000012
-2 4.44 5 7 -3 -2.5599999999999996
-20 10.3 2 22 18 -11.7
+##################################################
+# Clean up the funtions / mock tables at the end #
+##################################################
 
-# Drop the functions
 statement ok
 drop function add;
 
@@ -392,7 +415,6 @@ drop function add_named_wrapper;
 statement ok
 drop function type_match;
 
-# Drop the mock table
 statement ok
 drop table t1;