添加更多逻辑以获取匿名消息后，sql开始运行缓慢，因此我为字段添加了复合索引：
alter table dialog add index anonym_ftr1(from_anonym_id, to_anonym_id, removed_uid1);
alter table dialog add index anonym_tfr2(to_anonym_id, from_anonym_id, removed_uid2);
SQL查询如下所示：
SELECT
dialog.id AS dialog_id,
users.id AS uid,
users.purchases,
name,
number,
avatar,
avatar_date,
driver,
vehicle,
messages.id AS mid,
uid2,
mid2,
messages.msg,
messages.removed,
messages.from_anonym_id,
messages.to_anonym_id,
(messages.date DIV 1000) AS date,
from_id = 1 AS outbox,
(
SELECT
COUNT(*)
FROM
messages FORCE KEY (idx_ufsr)
WHERE
from_id = uid && user_id = 1 && read_state = 0 && removed = 0
) AS not_read
FROM
dialog,
messages,
users
WHERE
messages.id = dialog.mid && users.removed != 1 && (
(
uid1 = 1 && users.id = uid2 && removed_uid1 = 0
) || (
users.id = uid1 && uid2 = 1 && removed_uid2 = 0
)
OR (
dialog.from_anonym_id = 3
AND users.anonym_id = dialog.to_anonym_id && removed_uid1 = 0
)
OR (
dialog.to_anonym_id = 3
AND users.anonym_id = dialog.from_anonym_id && removed_uid2 = 0
)
)
ORDER BY
dialog.mid DESC
LIMIT 0,21
显示对话框中的索引：
+--------+------------+----------------+--------------+----------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| Table | Non_unique | Key_name | Seq_in_index | Column_name | Collation | Cardinality | Sub_part | Packed | Null | Index_type | Comment | Index_comment |
+--------+------------+----------------+--------------+----------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| dialog | 0 | PRIMARY | 1 | id | A | 5690715 | NULL | NULL | | BTREE | | |
| dialog | 1 | uid1 | 1 | uid1 | A | 379381 | NULL | NULL | | BTREE | | |
| dialog | 1 | uid1 | 2 | uid2 | A | 5690715 | NULL | NULL | | BTREE | | |
| dialog | 1 | uid2 | 1 | uid2 | A | 1896905 | NULL | NULL | | BTREE | | |
| dialog | 1 | idx_mid | 1 | mid | A | 5690715 | NULL | NULL | | BTREE | | |
| dialog | 1 | from_anonym_id | 1 | from_anonym_id | A | 18 | NULL | NULL | | BTREE | | |
| dialog | 1 | to_anonym_id | 1 | to_anonym_id | A | 18 | NULL | NULL | | BTREE | | |
| dialog | 1 | mid_uid | 1 | mid | A | 5690715 | NULL | NULL | | BTREE | | |
| dialog | 1 | mid_uid | 2 | uid2 | A | 5690715 | NULL | NULL | | BTREE | | |
| dialog | 1 | mid_uid | 3 | uid1 | A | 5690715 | NULL | NULL | | BTREE | | |
| dialog | 1 | mid2 | 1 | mid2 | A | 5690715 | NULL | NULL | | BTREE | | |
| dialog | 1 | uid_removed | 1 | uid1 | A | 18 | NULL | NULL | | BTREE | | |
| dialog | 1 | uid_removed | 2 | uid2 | A | 5690715 | NULL | NULL | | BTREE | | |
| dialog | 1 | uid_removed | 3 | removed_uid1 | A | 5690715 | NULL | NULL | | BTREE | | |
| dialog | 1 | uid_removed | 4 | removed_uid2 | A | 5690715 | NULL | NULL | | BTREE | | |
| dialog | 1 | anonym_ftr1 | 1 | from_anonym_id | A | 200 | NULL | NULL | | BTREE | | |
| dialog | 1 | anonym_ftr1 | 2 | to_anonym_id | A | 200 | NULL | NULL | | BTREE | | |
| dialog | 1 | anonym_ftr1 | 3 | removed_uid1 | A | 200 | NULL | NULL | | BTREE | | |
| dialog | 1 | anonym_tfr2 | 1 | to_anonym_id | A | 200 | NULL | NULL | | BTREE | | |
| dialog | 1 | anonym_tfr2 | 2 | from_anonym_id | A | 200 | NULL | NULL | | BTREE | | |
| dialog | 1 | anonym_tfr2 | 3 | removed_uid2 | A | 200 | NULL | NULL | | BTREE | | |
+--------+------------+----------------+--------------+----------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
解释延伸：
*************************** 1. row ***************************
id: 1
select_type: PRIMARY
table: dialog
type: index_merge
possible_keys: uid1,uid2,idx_mid,from_anonym_id,to_anonym_id,mid_uid,uid_removed,anonym_ftr1,anonym_tfr2
key: uid1,uid2,from_anonym_id,to_anonym_id
key_len: 4,4,4,4
ref: NULL
rows: 50
filtered: 100.00
Extra: Using sort_union(uid1,uid2,from_anonym_id,to_anonym_id); Using where; Using temporary; Using filesort
*************************** 2. row ***************************
id: 1
select_type: PRIMARY
table: users
type: range
possible_keys: PRIMARY,removed
key: removed
key_len: 1
ref: NULL
rows: 398596
filtered: 100.00
Extra: Using where; Using join buffer
*************************** 3. row ***************************
id: 1
select_type: PRIMARY
table: messages
type: eq_ref
possible_keys: PRIMARY
key: PRIMARY
key_len: 4
ref: anumbers.dialog.mid
rows: 1
filtered: 100.00
Extra:
*************************** 4. row ***************************
id: 2
select_type: DEPENDENT SUBQUERY
table: messages
type: ref
possible_keys: idx_ufsr
key: idx_ufsr
key_len: 10
ref: const,func,const,const
rows: 3
filtered: 100.00
Extra: Using index
4 rows in set, 2 warnings (0.00 sec)
我对联盟有一些想法，但是在中途排序会有问题。 此外，可以拆分2个查询并对它们进行排序。
来这里了解一下，如果有可能更快地运行1个查询？ 请帮我解决一个问题。真正帮助解决这个问题的人 - 将获得批准并指出等等。
PS抱歉，我不知道如何使用索引调整表格。