MDEV-24943: Implement FILTER clause support for aggregate functions

Aggregates lacked the SQL-standard FILTER clause, forcing CASE-based workarounds that reduced readability across (sum, avg, count, …).

This update introduces the ability to specify a FILTER clause for aggregate functions,
allowing for more granular control over which rows are included in the aggregation.
Also, improves standards compliance and makes queries clearer and more readable.

The FILTER(WHERE ...) condition may contain any expression allowed in regular WHERE clauses,
except subqueries, window functions, and outer references.

Author: KhaledR57

mysql-test/main/aggregates-filter.result

@@ -0,0 +1,329 @@
+#
+# Setup test table
+#
+#
+# Basic aggregates with WHERE
+#
+avg_result	sum_result	count_result	min_result	max_result	group_concat_result	bit_and_result	bit_or_result	bit_xor_result	json_arrayagg_result	json_objectagg_result	std_result	stddev_result	stddev_pop_result	stddev_samp_result	variance_result	var_pop_result	var_samp_result
+137.5000	550	3	75	200	Novel	0	15	6	["Phone","Tablet",null,null,"Laptop",null,null,null]	{"item2":"Tablet", "item5":"Laptop"}	44.7214	41.4578	0.0000	54.1987	1100.0000	4000.0000	312.5000
+#
+# Aggregates with DISTINCT and WHERE
+#
+avg_result	sum_result	count_result	min_result	max_result	group_concat_result
+175.0000	525	1	75	200	Laptop,Phone,Tablet
+#
+# Aggregates with complex WHERE conditions (AND)
+#
+avg_result	sum_result	count_result	min_result	max_result	group_concat_result	bit_and_result	bit_or_result	bit_xor_result	json_arrayagg_result	std_result	variance_result
+183.3333	250	5	75	200	Laptop,Phone,Tablet	0	15	9	[null,"Tablet",null,null,"Laptop",null,null,null]	41.4578	742.1875
+#
+# Aggregates with WHERE IN and OR
+#
+avg_result	sum_result	count_result	min_result	max_result	group_concat_result	bit_and_result	bit_or_result	bit_xor_result	json_arrayagg_result	std_result	variance_result
+137.5000	825	6	50	200	Guide,Novel,Pants,Phone,Shirt,Textbook	0	15	14	["Phone","Tablet",null,null,"Laptop","Shirt","Pants",null]	55.0973	2447.9167
+#
+# Aggregates with WHERE BETWEEN
+#
+avg_result	sum_result	count_result	min_result	max_result	group_concat_result	bit_and_result	bit_or_result	bit_xor_result	json_arrayagg_result	std_result	variance_result
+150.0000	625	5	75	150	Laptop,Pants,Phone,Shirt,Tablet	0	15	1	["Phone","Tablet",null,"Textbook","Laptop",null,null,"Guide"]	20.4124	555.5556
+#
+# Aggregates with WHERE LIKE
+#
+avg_result	sum_result	count_result	min_result	max_result	group_concat_result	bit_and_result	bit_or_result	bit_xor_result	json_arrayagg_result	std_result	variance_result
+131.2500	250	3	100	100	Guide,Novel,Textbook	6	15	5	[null,"Tablet",null,null,null,"Shirt",null,null]	12.5000	2447.9167
+#
+# Basic aggregates with GROUP BY
+#
+category	avg_result	sum_result	count_result	min_result	max_result	group_concat_result	bit_and_result	bit_or_result	bit_xor_result	json_arrayagg_result	std_result	variance_result
+Books	200.0000	400	2	NULL	200	NULL	7	15	8	[null,null,null]	0.0000	0.0000
+Clothing	75.0000	NULL	2	75	75	NULL	0	0	0	[null,null]	NULL	NULL
+Electronics	116.6667	150	3	100	100	Laptop,Phone,Tablet	0	15	1	[null,"Tablet","Laptop"]	23.5702	555.5556
+#
+# Aggregates with GROUP BY and WHERE IS NOT NULL
+#
+category	avg_result	sum_result	count_result	min_result	max_result	group_concat_result	bit_and_result	bit_or_result	bit_xor_result	json_arrayagg_result	std_result	variance_result
+Books	200.0000	400	2	200	200	Guide,Novel,Textbook	7	15	8	["Novel","Textbook","Guide"]	0.0000	0.0000
+Clothing	62.5000	75	2	75	75	Pants,Shirt	0	15	15	["Shirt","Pants"]	12.5000	156.2500
+Electronics	116.6667	350	3	100	150	Laptop,Phone,Tablet	0	15	1	["Phone","Tablet","Laptop"]	23.5702	555.5556
+#
+# Aggregates with multiple GROUP BY columns
+#
+category	status	avg_result	sum_result	count_result	min_result	max_result	group_concat_result	bit_and_result	bit_or_result	bit_xor_result	json_arrayagg_result	std_result	variance_result
+Books	active	200.0000	400	2	200	200	Guide,Textbook	7	15	8	["Textbook","Guide"]	0.0000	0.0000
+Books	inactive	NULL	NULL	1	NULL	NULL	Novel	18446744073709551615	0	0	["Novel"]	NULL	NULL
+Clothing	NULL	50.0000	50	1	NULL	50	Shirt	3	0	3	["Shirt"]	0.0000	0.0000
+Clothing	active	75.0000	75	1	75	75	Pants	12	12	12	["Pants"]	0.0000	0.0000
+Electronics	active	116.6667	350	3	100	150	Laptop,Phone,Tablet	0	15	9	["Phone","Tablet","Laptop"]	23.5702	555.5556
+#
+# GROUP BY with HAVING on aggregates
+#
+category	avg_result	sum_result	count_result	min_result	max_result
+Books	200.0000	400	2	NULL	200
+#
+# HAVING with alias
+#
+category	avg_value	sum_value	count_value	min_value	max_value
+Books	200.0000	400	2	NULL	200
+#
+# WHERE + GROUP BY + HAVING
+#
+category	avg_result	sum_result	count_result	min_result	max_result
+Books	200.0000	400	2	NULL	200
+Electronics	116.6667	150	3	100	100
+#
+# WHERE + multiple aggregates with FILTER
+#
+category	total_avg	active_avg	total_sum	active_sum	total_count	active_count	total_min	active_min	total_max	active_max
+Books	200.0000	200.0000	400	NULL	2	2	NULL	NULL	200	200
+Clothing	62.5000	75.0000	NULL	NULL	2	2	NULL	NULL	75	75
+Electronics	116.6667	116.6667	150	350	3	1	100	100	150	150
+#
+# WHERE + FILTER + HAVING
+#
+category	active_avg	active_sum	active_count
+Books	200.0000	400	0
+Clothing	75.0000	NULL	0
+Electronics	116.6667	150	3
+#
+# WHERE + window function (OVER)
+#
+category	cat_avg	cat_sum	cat_count	cat_min	cat_max
+Books	200.0000	400	2	NULL	200
+Books	200.0000	400	2	NULL	200
+Books	200.0000	400	2	NULL	200
+Clothing	75.0000	NULL	2	75	75
+Clothing	75.0000	NULL	2	75	75
+Electronics	116.6667	150	3	100	100
+Electronics	116.6667	150	3	100	100
+Electronics	116.6667	150	3	100	100
+#
+# WHERE + window function with ORDER BY
+#
+id	running_avg	running_sum	running_count	running_min	running_max
+1	100.0000	100	1	100	100
+2	125.0000	250	2	100	150
+3	125.0000	250	3	100	150
+4	150.0000	450	4	100	200
+5	137.5000	550	5	100	200
+6	120.0000	550	6	100	200
+7	112.5000	625	7	75	200
+8	125.0000	825	8	75	200
+#
+# WHERE + window function with frame specification
+#
+id	windowed_avg	windowed_sum	windowed_count	windowed_min	windowed_max
+1	125.0000	150	2	100	NULL
+2	125.0000	150	2	100	NULL
+3	175.0000	350	2	150	200
+4	150.0000	200	2	100	200
+5	150.0000	200	3	100	200
+6	87.5000	NULL	3	75	100
+7	137.5000	200	3	75	200
+8	137.5000	200	2	75	200
+#
+# Subquery with aggregates in WHERE
+#
+id	category	status	value	price	amount	name	key_name	value_col	bit_value	geom
+3	Books	inactive	NULL	19.99	19.99	Novel	item3	Novel	NULL	POINT(3 3)
+4	Books	active	200	29.99	29.99	Textbook	item4	Textbook	7	POINT(4 4)
+8	Books	active	200	24.99	24.99	Guide	item8	Guide	15	POINT(8 8)
+#
+# Aggregates with GROUP BY and ORDER BY
+#
+category	avg_val	sum_val	count_val	min_val	max_val
+Books	200.0000	400	2	NULL	200
+Electronics	116.6667	150	3	100	100
+Clothing	75.0000	NULL	2	75	75
+#
+# Complete query with WHERE, GROUP BY, HAVING, ORDER BY
+#
+category	avg_val	sum_val	count_val	min_val	max_val
+Books	200.0000	400	2	NULL	200
+Electronics	116.6667	150	3	100	100
+#
+# Aggregates with GROUP BY, ORDER BY, and LIMIT
+#
+category	avg_val	sum_val	count_val
+Books	200.0000	400	2
+Electronics	116.6667	150	3
+#
+# Aggregates with ROLLUP
+#
+category	status	avg_val	sum_val	count_val	min_val	max_val
+Books	active	200.0000	400	2	NULL	200
+Books	inactive	NULL	NULL	0	NULL	NULL
+Books	NULL	200.0000	400	2	NULL	200
+Clothing	NULL	NULL	NULL	1	NULL	50
+Clothing	active	75.0000	NULL	1	75	75
+Clothing	NULL	75.0000	NULL	2	75	75
+Electronics	active	116.6667	150	3	100	100
+Electronics	NULL	116.6667	150	3	100	100
+NULL	NULL	137.5000	550	7	75	200
+#
+# Most complete query with all clauses and aggregates
+#
+category	status	avg_value	sum_value	count_value	min_value	max_value	avg_amount	sum_amount	distinct_value_avg	distinct_value_sum	distinct_value_count	group_concat_result	bit_and_result	bit_or_result	bit_xor_result	json_arrayagg_result	std_result	variance_result
+Books	active	200.0000	400	2	NULL	200	NULL	54.98	200.0000	200	1	Guide	7	15	0	[null,null]	NULL	0.0000
+Electronics	active	100.0000	350	1	100	150	333.326667	999.98	100.0000	NULL	2	Laptop	0	15	7	[null,null,null]	23.5702	555.5556
+Clothing	active	75.0000	75	0	75	75	NULL	NULL	75.0000	NULL	1	Pants	18446744073709551615	0	0	[null]	0.0000	NULL
+#
+# COUNT with column names (not just *)
+#
+count_value	count_category	count_name
+6	3	3
+#
+# COUNT DISTINCT with multiple columns
+#
+count_distinct
+3
+#
+# FILTER on indexed column
+#
+# Create index on status column
+avg_result	sum_result	count_result
+137.5000	350	3
+#
+# Empty result set (all rows filtered out)
+#
+avg_result	sum_result	count_result	min_result	max_result
+NULL	NULL	0	NULL	NULL
+#
+# FILTER with NULL conditions
+#
+avg_result	count_result	sum_result
+NULL	0	NULL
+#
+# FILTER with JOINs
+#
+# Create second table for JOIN test
+category	avg_result	sum_result	count_result
+Books	200.0000	200	2
+Electronics	150.0000	NULL	0
+#
+# FILTER with window functions - different partitions
+#
+category	status	cat_avg	status_avg	cat_count	status_count
+Books	active	200.0000	183.3333	2	4
+Books	active	200.0000	183.3333	2	4
+Books	inactive	200.0000	NULL	2	0
+Clothing	NULL	75.0000	NULL	2	0
+Clothing	active	75.0000	183.3333	2	4
+Electronics	active	116.6667	183.3333	3	4
+Electronics	active	116.6667	183.3333	3	4
+Electronics	active	116.6667	183.3333	3	4
+#
+# FILTER with empty table
+#
+# Create empty table
+avg_result	sum_result	count_result	min_result	max_result
+NULL	NULL	0	NULL	NULL
+#
+# FILTER with CASE in WHERE clause
+#
+avg_result	sum_result	count_result
+137.5000	550	3
+#
+# ORDER BY with FILTER aggregates
+#
+category	avg_val	sum_val
+Books	200.0000	400
+Electronics	116.6667	150
+Clothing	75.0000	NULL
+#
+# Aggregates on id column itself with FILTER
+#
+avg_id	sum_id	count_id	min_id	max_id	distinct_count_id
+4.5000	14	3	1	8	5
+#
+# Aggregates on id with GROUP BY and FILTER
+#
+category	avg_id	sum_id	count_id	min_id	max_id
+Books	6.0000	12	2	NULL	8
+Clothing	6.5000	7	0	7	7
+Electronics	5.0000	8	1	1	5
+#
+# ========================================
+# ERROR CASES SECTION
+# ========================================
+#
+#
+# FILTER with Non-Aggregate Window Functions (Ranking)
+#
+# FILTER with RANK (should error - FILTER only works with aggregates)
+ERROR 42000: You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near 'FILTER (WHERE status = 'active') OVER (ORDER BY value) FROM test_aggregates' at line 1
+#
+# FILTER with Non-Aggregate Window Functions (Navigation)
+#
+# FILTER with LAG (should error - FILTER only works with aggregates)
+ERROR 42000: You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near 'FILTER (WHERE status = 'active') OVER (ORDER BY id) FROM test_aggregates' at line 1
+#
+# Wrong Order - OVER before FILTER
+#
+# OVER clause before FILTER clause (should error - correct order is FILTER then OVER)
+ERROR 42000: You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near '(WHERE status = 'active') FROM test_aggregates' at line 1
+#
+# Non-Aggregate Scalar Functions
+#
+# Regular scalar function (should error - FILTER only for aggregates)
+ERROR 42000: You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near '(WHERE status = 'active') FROM test_aggregates' at line 1
+#
+# Plain Column or Expression
+#
+# FILTER on plain column (should error - not a function)
+ERROR 42000: You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near '(WHERE status = 'active') FROM test_aggregates' at line 1
+#
+# Empty or Invalid Clause
+#
+# Empty FILTER clause (should error - WHERE condition required)
+ERROR 42000: You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near ') FROM test_aggregates' at line 1
+#
+# Nested FILTER Clauses
+#
+# Nested FILTER clauses (should error - cannot nest FILTER)
+ERROR HY000: Invalid use of group function
+#
+# Invalid WHERE Syntax
+#
+# Multiple WHERE keywords (should error)
+ERROR 42000: You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near 'WHERE value > 100) FROM test_aggregates' at line 1
+#
+# Column Aliases
+#
+# Using column alias in FILTER WHERE (should error - alias not available)
+ERROR 42S22: Unknown column 'val' in 'SELECT'
+#
+# Subqueries in WHERE Condition
+#
+# Scalar subquery in FILTER WHERE (should error - subqueries not allowed)
+ERROR HY000: Incorrect usage of subqueries in FILTER clause and 
+# IN subquery in FILTER WHERE (should error - subqueries not allowed)
+ERROR HY000: Incorrect usage of subqueries in FILTER clause and Item_in_subselect::fix_fields
+#
+# FILTER with Window Functions in WHERE Condition
+#
+# Window function in FILTER WHERE condition (should error - not allowed)
+ERROR HY000: Incorrect usage of window function and FILTER clause
+#
+# FILTER with aggregate functions in WHERE clause (should expect error)
+#
+# This should error - aggregate in WHERE clause
+ERROR HY000: Invalid use of group function
+#
+# FILTER with error cases - syntax errors
+#
+# Missing WHERE keyword (should error)
+ERROR 42000: You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near 'status = 'active') FROM test_aggregates' at line 1
+# Missing parentheses (should error)
+ERROR 42000: You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near 'WHERE status = 'active' FROM test_aggregates' at line 1
+#
+# Illegal use of FILTER with non-aggregate UDFs
+#
+# Create a non-aggregate UDF for testing
+# FILTER clause with non-aggregate UDF should fail
+ERROR 42000: You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near '(WHERE status = 'active') FROM test_aggregates' at line 1
+#
+# CLEANUP
+#
+#
+# End of aggregates FILTER test
+#