database/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-08-08 11:22:35 +03:00
Code Activity

MDEV-30048 Prefix keys for CHAR work differently for MyISAM vs InnoDB

Browse Source 
Also fixes: MDEV-30050 Inconsistent results of DISTINCT with NOPAD

Problem:

Key segments for CHAR columns where compared using strnncollsp()
for engines MyISAM and Aria.

This did not work correct in case if the engine applyied trailing
space compression.

Fix:

Replacing ha_compare_text() calls to new functions:

- ha_compare_char_varying()
- ha_compare_char_fixed()
- ha_compare_word()
- ha_compare_word_prefix()
- ha_compare_word_or_prefix()

The code branch corresponding to comparison of CHAR column keys
(HA_KEYTYPE_TEXT segment type) now uses ha_compare_char_fixed()
which calls strnncollsp_nchars().

This patch does not change the behavior for the rest of the code:
- comparison of VARCHAR/TEXT column keys
  (HA_KEYTYPE_VARTEXT1, HA_KEYTYPE_VARTEXT2 segments types)
- comparison in the fulltext code
This commit is contained in:
Alexander Barkov
2023-04-07 12:15:41 +04:00
parent 09e237088c
commit df72c57d6f
26 changed files with 634 additions and 102 deletions
Download Patch File Download Diff File Expand all files Collapse all files

131
include/my_compare.h
View File

@@ -110,8 +110,135 @@ static inline void set_rec_bits(uint16 bits, uchar *ptr, uchar ofs, uint len)
#define clr_rec_bits(bit_ptr, bit_ofs, bit_len) \
set_rec_bits(0, bit_ptr, bit_ofs, bit_len)
extern int ha_compare_text(CHARSET_INFO *, const uchar *, size_t,
const uchar *, size_t , my_bool);
/*
Compare two VARCHAR values.
@param charset_info - The character set and collation
@param a - The pointer to the first string
@param a_length - The length of the first string
@param b - The pointer to the second string
@param b_length - The length of the second string
@param b_is_prefix - Whether "b" is a prefix of "a",
e.g. in a prefix key (partial length key).
@returns - The result of comparison
- If "b_is_prefix" is FALSE, then the two strings are compared
taking into account the PAD SPACE/NO PAD attribute of the collation.
- If "b_is_prefix" is TRUE, then trailing spaces are compared in NO PAD style.
This is done e.g. when we compare a column value to its prefix key value
(the value of "a" to the value of "key_a"):
CREATE TABLE t1 (a VARCHAR(10), KEY(key_a(5));
*/
static inline int ha_compare_char_varying(CHARSET_INFO *charset_info,
const uchar *a, size_t a_length,
const uchar *b, size_t b_length,
my_bool b_is_prefix)
{
if (!b_is_prefix)
return charset_info->coll->strnncollsp(charset_info, a, a_length,
b, b_length);
return charset_info->coll->strnncoll(charset_info,
a, a_length,
b, b_length, TRUE/*prefix*/);
}
/*
Compare two CHAR values of the same declared character length,
e.g. CHAR(5) to CHAR(5).
@param charset_info - The character set and collation
@param a - The pointer to the first string
@param a_length - The length of the first string
@param b - The pointer to the second string
@param b_length - The length of the second string
@param nchars - The declared length (in characters)
@param b_is_prefix - Whether "b" is a prefix of "a",
e.g. in a prefix key (partial length key).
@returns - The result of comparison
- If "b_is_prefix" is FALSE, then the two strings are compared
taking into account the PAD SPACE/NO PAD attribute of the collation.
Additionally, this function assumes that the underlying storage could
optionally apply trailing space compression, so values can come into this
comparison function in different states:
- all trailing spaces removed
- some trailing spaced removed
- no trailing spaces removed (exactly "nchars" characters on the two sides)
This function virtually reconstructs trailing spaces up to the defined
length specified in "nchars".
If either of the sides have more than "nchar" characters,
then only leftmost "nchar" characters are compared.
- If "b_is_prefix" is TRUE, then trailing spaces are compared in NO PAD style.
This is done e.g. when we compare a column value to its prefix key value
(the value of "a" to the value of "key_a"):
CREATE TABLE t1 (a CHAR(10), KEY(key_a(5));
*/
static inline int ha_compare_char_fixed(CHARSET_INFO *charset_info,
const uchar *a, size_t a_length,
const uchar *b, size_t b_length,
size_t nchars,
my_bool b_is_prefix)
{
if (!b_is_prefix)
return charset_info->coll->strnncollsp_nchars(charset_info,
a, a_length,
b, b_length,
nchars,
MY_STRNNCOLLSP_NCHARS_EMULATE_TRIMMED_TRAILING_SPACES);
return charset_info->coll->strnncoll(charset_info,
a, a_length,
b, b_length, TRUE/*prefix*/);
}
/*
A function to compare words of a text.
This is a common operation in full-text search:
SELECT MATCH (title) AGAINST ('word') FROM t1;
*/
static inline int ha_compare_word(CHARSET_INFO *charset_info,
const uchar *a, size_t a_length,
const uchar *b, size_t b_length)
{
return charset_info->coll->strnncollsp(charset_info,
a, a_length,
b, b_length);
}
/*
A function to compare a word of a text to a word prefix.
This is a common operation in full-text search:
SELECT MATCH (title) AGAINST ('wor*' IN BOOLEAN MODE) FROM t1;
*/
static inline int ha_compare_word_prefix(CHARSET_INFO *charset_info,
const uchar *a, size_t a_length,
const uchar *b, size_t b_length)
{
return charset_info->coll->strnncoll(charset_info,
a, a_length,
b, b_length,
TRUE/*b_is_prefix*/);
}
/*
Compare words (full match or prefix match), e.g. for full-text search.
*/
static inline int ha_compare_word_or_prefix(CHARSET_INFO *charset_info,
const uchar *a, size_t a_length,
const uchar *b, size_t b_length,
my_bool b_is_prefix)
{
if (!b_is_prefix)
return ha_compare_word(charset_info, a, a_length, b, b_length);
return ha_compare_word_prefix(charset_info, a, a_length, b, b_length);
}
extern int ha_key_cmp(HA_KEYSEG *keyseg, const uchar *a,
const uchar *b, uint key_length, uint nextflag,
uint *diff_pos);