[feature](information_schema)add metadata_name_ids for quickly get catlogs,db,table and add profiling table in order to Compatible with mysql (#22702)

add information_schema.metadata_name_idsfor quickly get catlogs,db,table.

1. table  struct :   
```mysql
mysql> desc  internal.information_schema.metadata_name_ids;
+---------------+--------------+------+-------+---------+-------+
| Field         | Type         | Null | Key   | Default | Extra |
+---------------+--------------+------+-------+---------+-------+
| CATALOG_ID    | BIGINT       | Yes  | false | NULL    |       |
| CATALOG_NAME  | VARCHAR(512) | Yes  | false | NULL    |       |
| DATABASE_ID   | BIGINT       | Yes  | false | NULL    |       |
| DATABASE_NAME | VARCHAR(64)  | Yes  | false | NULL    |       |
| TABLE_ID      | BIGINT       | Yes  | false | NULL    |       |
| TABLE_NAME    | VARCHAR(64)  | Yes  | false | NULL    |       |
+---------------+--------------+------+-------+---------+-------+
6 rows in set (0.00 sec) 


mysql> select * from internal.information_schema.metadata_name_ids where CATALOG_NAME="hive1" limit 1 \G;
*************************** 1. row ***************************
   CATALOG_ID: 113008
 CATALOG_NAME: hive1
  DATABASE_ID: 113042
DATABASE_NAME: ssb1_parquet
     TABLE_ID: 114009
   TABLE_NAME: dates
1 row in set (0.07 sec)
```

2. when you create / drop catalog , need not refresh catalog . 
```mysql
mysql> select count(*) from internal.information_schema.metadata_name_ids\G; 
*************************** 1. row ***************************
count(*): 21301
1 row in set (0.34 sec)


mysql> drop catalog hive2;
Query OK, 0 rows affected (0.01 sec)

mysql> select count(*) from internal.information_schema.metadata_name_ids\G; 
*************************** 1. row ***************************
count(*): 10665
1 row in set (0.04 sec) 


mysql> create catalog hive3 ... 
mysql> select count(*) from internal.information_schema.metadata_name_ids\G;                                                                        
*************************** 1. row ***************************
count(*): 21301
1 row in set (0.32 sec)
```

3. create / drop table , need not refresh catalog .  
```mysql
mysql> CREATE TABLE IF NOT EXISTS demo.example_tbl ... ;


mysql> select count(*) from internal.information_schema.metadata_name_ids\G; 
*************************** 1. row ***************************
count(*): 10666
1 row in set (0.04 sec)

mysql> drop table demo.example_tbl;
Query OK, 0 rows affected (0.01 sec)

mysql> select count(*) from internal.information_schema.metadata_name_ids\G; 
*************************** 1. row ***************************
count(*): 10665
1 row in set (0.04 sec) 

```

4. you can set query time , prevent queries from taking too long . 
```

fe.conf :  query_metadata_name_ids_timeout 

the time used to obtain all tables in one database

```
5. add information_schema.profiling in order to Compatible with  mysql

```mysql
mysql> select * from information_schema.profiling;
Empty set (0.07 sec)

mysql> set profiling=1;                                                                                 
Query OK, 0 rows affected (0.01 sec)
```
This commit is contained in:
daidai
2023-08-31 21:22:26 +08:00
committed by GitHub
parent 6fe2418cfc
commit e680d42fe7
19 changed files with 743 additions and 52 deletions

View File

@ -2136,4 +2136,11 @@ public class Config extends ConfigBase {
@ConfField
public static int table_stats_health_threshold = 80;
@ConfField(mutable = true, masterOnly = false, description = {
"查询information_schema.metadata_name_ids表时,获取一个数据库中所有表用的时间",
"When querying the information_schema.metadata_name_ids table,"
+ " the time used to obtain all tables in one database"
})
public static long query_metadata_name_ids_timeout = 3;
}