Commit 59ab9a08 authored by Georgi Kodinov's avatar Georgi Kodinov

Bug#37830 : ORDER BY ASC/DESC - no difference

            
Range scan in descending order for c <= <col> <= c type of
ranges was ignoring the DESC flag.
However some engines like InnoDB have the primary key parts 
as a suffix for every secondary key.
When such primary key suffix is used for ordering ignoring 
the DESC is not valid.
But we generally would like to do this because it's faster.
      
Fixed by performing only reverse scan if the primary key is used.
Removed some dead code in the process.
parent 52f510ef
...@@ -1246,4 +1246,19 @@ set global innodb_autoextend_increment=@my_innodb_autoextend_increment; ...@@ -1246,4 +1246,19 @@ set global innodb_autoextend_increment=@my_innodb_autoextend_increment;
set @my_innodb_commit_concurrency=@@global.innodb_commit_concurrency; set @my_innodb_commit_concurrency=@@global.innodb_commit_concurrency;
set global innodb_commit_concurrency=0; set global innodb_commit_concurrency=0;
set global innodb_commit_concurrency=@my_innodb_commit_concurrency; set global innodb_commit_concurrency=@my_innodb_commit_concurrency;
CREATE TABLE t1 (a int, b int, c int, PRIMARY KEY (a), KEY t1_b (b))
ENGINE=InnoDB;
INSERT INTO t1 (a,b,c) VALUES (1,1,1), (2,1,1), (3,1,1), (4,1,1);
INSERT INTO t1 (a,b,c) SELECT a+4,b,c FROM t1;
EXPLAIN SELECT a, b, c FROM t1 WHERE b = 1 ORDER BY a DESC LIMIT 5;
id select_type table type possible_keys key key_len ref rows Extra
1 SIMPLE t1 range t1_b t1_b 5 NULL 4 Using where
SELECT a, b, c FROM t1 WHERE b = 1 ORDER BY a DESC LIMIT 5;
a b c
8 1 1
7 1 1
6 1 1
5 1 1
4 1 1
DROP TABLE t1;
End of 5.0 tests End of 5.0 tests
...@@ -996,4 +996,22 @@ set @my_innodb_commit_concurrency=@@global.innodb_commit_concurrency; ...@@ -996,4 +996,22 @@ set @my_innodb_commit_concurrency=@@global.innodb_commit_concurrency;
set global innodb_commit_concurrency=0; set global innodb_commit_concurrency=0;
set global innodb_commit_concurrency=@my_innodb_commit_concurrency; set global innodb_commit_concurrency=@my_innodb_commit_concurrency;
#
# Bug #37830: ORDER BY ASC/DESC - no difference
#
CREATE TABLE t1 (a int, b int, c int, PRIMARY KEY (a), KEY t1_b (b))
ENGINE=InnoDB;
INSERT INTO t1 (a,b,c) VALUES (1,1,1), (2,1,1), (3,1,1), (4,1,1);
INSERT INTO t1 (a,b,c) SELECT a+4,b,c FROM t1;
# should be range access
EXPLAIN SELECT a, b, c FROM t1 WHERE b = 1 ORDER BY a DESC LIMIT 5;
# should produce '8 7 6 5 4' for a
SELECT a, b, c FROM t1 WHERE b = 1 ORDER BY a DESC LIMIT 5;
DROP TABLE t1;
--echo End of 5.0 tests --echo End of 5.0 tests
...@@ -7094,7 +7094,8 @@ bool QUICK_RANGE_SELECT::row_in_ranges() ...@@ -7094,7 +7094,8 @@ bool QUICK_RANGE_SELECT::row_in_ranges()
QUICK_SELECT_DESC::QUICK_SELECT_DESC(QUICK_RANGE_SELECT *q, QUICK_SELECT_DESC::QUICK_SELECT_DESC(QUICK_RANGE_SELECT *q,
uint used_key_parts_arg) uint used_key_parts_arg)
: QUICK_RANGE_SELECT(*q), rev_it(rev_ranges) : QUICK_RANGE_SELECT(*q), rev_it(rev_ranges),
used_key_parts (used_key_parts_arg)
{ {
QUICK_RANGE *r; QUICK_RANGE *r;
...@@ -7136,10 +7137,11 @@ int QUICK_SELECT_DESC::get_next() ...@@ -7136,10 +7137,11 @@ int QUICK_SELECT_DESC::get_next()
int result; int result;
if (last_range) if (last_range)
{ // Already read through key { // Already read through key
result = ((last_range->flag & EQ_RANGE) result = ((last_range->flag & EQ_RANGE &&
? file->index_next_same(record, (byte*) last_range->min_key, used_key_parts <= head->key_info[index].key_parts) ?
last_range->min_length) : file->index_next_same(record, (byte*) last_range->min_key,
file->index_prev(record)); last_range->min_length) :
file->index_prev(record));
if (!result) if (!result)
{ {
if (cmp_prev(*rev_it.ref()) == 0) if (cmp_prev(*rev_it.ref()) == 0)
...@@ -7163,7 +7165,9 @@ int QUICK_SELECT_DESC::get_next() ...@@ -7163,7 +7165,9 @@ int QUICK_SELECT_DESC::get_next()
continue; continue;
} }
if (last_range->flag & EQ_RANGE) if (last_range->flag & EQ_RANGE &&
used_key_parts <= head->key_info[index].key_parts)
{ {
result= file->index_read(record, (byte*) last_range->max_key, result= file->index_read(record, (byte*) last_range->max_key,
last_range->max_length, HA_READ_KEY_EXACT); last_range->max_length, HA_READ_KEY_EXACT);
...@@ -7171,6 +7175,8 @@ int QUICK_SELECT_DESC::get_next() ...@@ -7171,6 +7175,8 @@ int QUICK_SELECT_DESC::get_next()
else else
{ {
DBUG_ASSERT(last_range->flag & NEAR_MAX || DBUG_ASSERT(last_range->flag & NEAR_MAX ||
(last_range->flag & EQ_RANGE &&
used_key_parts > head->key_info[index].key_parts) ||
range_reads_after_key(last_range)); range_reads_after_key(last_range));
result=file->index_read(record, (byte*) last_range->max_key, result=file->index_read(record, (byte*) last_range->max_key,
last_range->max_length, last_range->max_length,
...@@ -7268,54 +7274,6 @@ bool QUICK_SELECT_DESC::range_reads_after_key(QUICK_RANGE *range_arg) ...@@ -7268,54 +7274,6 @@ bool QUICK_SELECT_DESC::range_reads_after_key(QUICK_RANGE *range_arg)
} }
/* TRUE if we are reading over a key that may have a NULL value */
#ifdef NOT_USED
bool QUICK_SELECT_DESC::test_if_null_range(QUICK_RANGE *range_arg,
uint used_key_parts)
{
uint offset, end;
KEY_PART *key_part = key_parts,
*key_part_end= key_part+used_key_parts;
for (offset= 0, end = min(range_arg->min_length, range_arg->max_length) ;
offset < end && key_part != key_part_end ;
offset+= key_part++->store_length)
{
if (!memcmp((char*) range_arg->min_key+offset,
(char*) range_arg->max_key+offset,
key_part->store_length))
continue;
if (key_part->null_bit && range_arg->min_key[offset])
return 1; // min_key is null and max_key isn't
// Range doesn't cover NULL. This is ok if there is no more null parts
break;
}
/*
If the next min_range is > NULL, then we can use this, even if
it's a NULL key
Example: SELECT * FROM t1 WHERE a = 2 AND b >0 ORDER BY a DESC,b DESC;
*/
if (key_part != key_part_end && key_part->null_bit)
{
if (offset >= range_arg->min_length || range_arg->min_key[offset])
return 1; // Could be null
key_part++;
}
/*
If any of the key parts used in the ORDER BY could be NULL, we can't
use the key to sort the data.
*/
for (; key_part != key_part_end ; key_part++)
if (key_part->null_bit)
return 1; // Covers null part
return 0;
}
#endif
void QUICK_RANGE_SELECT::add_info_string(String *str) void QUICK_RANGE_SELECT::add_info_string(String *str)
{ {
KEY *key_info= head->key_info + index; KEY *key_info= head->key_info + index;
......
...@@ -667,12 +667,10 @@ public: ...@@ -667,12 +667,10 @@ public:
int get_type() { return QS_TYPE_RANGE_DESC; } int get_type() { return QS_TYPE_RANGE_DESC; }
private: private:
bool range_reads_after_key(QUICK_RANGE *range); bool range_reads_after_key(QUICK_RANGE *range);
#ifdef NOT_USED
bool test_if_null_range(QUICK_RANGE *range, uint used_key_parts);
#endif
int reset(void) { rev_it.rewind(); return QUICK_RANGE_SELECT::reset(); } int reset(void) { rev_it.rewind(); return QUICK_RANGE_SELECT::reset(); }
List<QUICK_RANGE> rev_ranges; List<QUICK_RANGE> rev_ranges;
List_iterator<QUICK_RANGE> rev_it; List_iterator<QUICK_RANGE> rev_it;
uint used_key_parts;
}; };
......
...@@ -12088,26 +12088,25 @@ part_of_refkey(TABLE *table,Field *field) ...@@ -12088,26 +12088,25 @@ part_of_refkey(TABLE *table,Field *field)
} }
/***************************************************************************** /**
Test if one can use the key to resolve ORDER BY Test if a key can be used to resolve ORDER BY
SYNOPSIS used_key_parts is set to correct key parts used if return value != 0
test_if_order_by_key() (On other cases, used_key_part may be changed).
order Sort order Note that the value may actually be greater than the number of index
table Table to sort key parts. This can happen for storage engines that have the primary
idx Index to check key parts as a suffix for every secondary key.
used_key_parts Return value for used key parts.
@param order Sort order
@param table Table to sort
NOTES @param idx Index to check
used_key_parts is set to correct key parts used if return value != 0 @param[out] used_key_parts Return value for used key parts.
(On other cases, used_key_part may be changed)
@return indication if the key can be used for sorting
RETURN @retval 1 key can be used for reading data in order.
1 key is ok. @retval 0 Key can't be used
0 Key can't be used @retval -1 Reverse read on the key can be used
-1 Reverse key can be used */
*****************************************************************************/
static int test_if_order_by_key(ORDER *order, TABLE *table, uint idx, static int test_if_order_by_key(ORDER *order, TABLE *table, uint idx,
uint *used_key_parts) uint *used_key_parts)
...@@ -12172,11 +12171,27 @@ static int test_if_order_by_key(ORDER *order, TABLE *table, uint idx, ...@@ -12172,11 +12171,27 @@ static int test_if_order_by_key(ORDER *order, TABLE *table, uint idx,
reverse=flag; // Remember if reverse reverse=flag; // Remember if reverse
key_part++; key_part++;
} }
*used_key_parts= on_primary_key ? table->key_info[idx].key_parts : if (on_primary_key)
(uint) (key_part - table->key_info[idx].key_part); {
if (reverse == -1 && !(table->file->index_flags(idx, *used_key_parts-1, 1) & uint used_key_parts_secondary= table->key_info[idx].key_parts;
HA_READ_PREV)) uint used_key_parts_pk=
reverse= 0; // Index can't be used (uint) (key_part - table->key_info[table->s->primary_key].key_part);
*used_key_parts= used_key_parts_pk + used_key_parts_secondary;
if (reverse == -1 &&
(!(table->file->index_flags(idx, used_key_parts_secondary - 1, 1) &
HA_READ_PREV) ||
!(table->file->index_flags(table->s->primary_key,
used_key_parts_pk - 1, 1) & HA_READ_PREV)))
reverse= 0; // Index can't be used
}
else
{
*used_key_parts= (uint) (key_part - table->key_info[idx].key_part);
if (reverse == -1 &&
!(table->file->index_flags(idx, *used_key_parts-1, 1) & HA_READ_PREV))
reverse= 0; // Index can't be used
}
DBUG_RETURN(reverse); DBUG_RETURN(reverse);
} }
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment