*: add sys schema, sys.SCHEMA_UNUSED_INDEXES view and sys.SCHEMA_INDEX_USAGE view
#22126
Conversation
|
/label status/WIP |
|
I'll add tests for this PR later. |
|
/label sig/planner |
|
/label component/statisitic |
|
These labels are not found |
|
/label component/statistic |
|
These labels are not found |
|
/label component/statistics |
rebelice
changed the title
sys schema and sys.SCHEMA_UNUSED_INDEXES viewsys schema and sys.SCHEMA_UNUSED_INDEXES view， sys.SCHEMA_INDEX_USAGE view
rebelice
changed the title
sys schema and sys.SCHEMA_UNUSED_INDEXES view， sys.SCHEMA_INDEX_USAGE viewsys schema, sys.SCHEMA_UNUSED_INDEXES view and sys.SCHEMA_INDEX_USAGE view
|
/unlabel status/WIP |
|
/cc @qw4990 |
infoschema/sysschema/tables.go
Outdated
| var pluginTable = make(map[string]func(autoid.Allocators, *model.TableInfo) (table.Table, error)) | ||
|
|
||
| // RegisterTable registers a new table into TiDB. | ||
| func RegisterTable(tableName, sql string, |
| } | ||
|
|
||
| const viewUnusedIndexes = "CREATE VIEW sys." + viewNameUnusedIndexes + | ||
| ` AS SELECT DISTINCT i.table_schema AS table_schema, i.table_name AS table_name, i.key_name AS index_name |
There was a problem hiding this comment.
Why use DISTINCT in these two SQLs?
There was a problem hiding this comment.
Consider a situation， table t has an index idx_ab(a, b). There will be two rows of data in the table information_schema.tidb_indexes because this index covers two columns. So there will be two lines of the same (table_schema, t, idx_ab). DISTINCT is to deal with this situation.
| @@ -1923,6 +1923,52 @@ func (d *ddl) RecoverTable(ctx sessionctx.Context, recoverInfo *RecoverInfo) (er | |||
| return errors.Trace(err) | |||
| } | |||
|
|
|||
| // BuildTableInfoFromCreateViewAST builds model.TableInfo from a CreateViewStmt statement. | |||
| func BuildTableInfoFromCreateViewAST(s *ast.CreateViewStmt) (*model.TableInfo, error) { | |||
There was a problem hiding this comment.
How about renaming it to BuildViewInfoFromAST to correspond to BuildTableInfoFromAST?
There was a problem hiding this comment.
I don’t think this is a good name, because this function does not build ViewInfo, but TableInfo.
infoschema/sysschema/const.go
Outdated
| @@ -0,0 +1,49 @@ | |||
| // Copyright 2016 PingCAP, Inc. | |||
There was a problem hiding this comment.
Please change 2016 to 2021.
infoschema/sysschema/init.go
Outdated
| @@ -0,0 +1,76 @@ | |||
| // Copyright 2016 PingCAP, Inc. | |||
infoschema/sysschema/tables.go
Outdated
| @@ -0,0 +1,101 @@ | |||
| // Copyright 2017 PingCAP, Inc. | |||
infoschema/sysschema/tables_test.go
Outdated
| @@ -0,0 +1,63 @@ | |||
| // Copyright 2016 PingCAP, Inc. | |||
meta/autoid/autoid.go
Outdated
| @@ -46,6 +46,7 @@ const ( | |||
| PerformanceSchemaDBID int64 = SystemSchemaIDFlag | 10000 | |||
| // MetricSchemaDBID is the metrics_schema schema id, it's exported for test. | |||
| MetricSchemaDBID int64 = SystemSchemaIDFlag | 20000 | |||
| SysSchemaDBID int64 = SystemSchemaIDFlag | 40000 | |||
There was a problem hiding this comment.
Please add some comments. Does the number 40000 is a magic number, just need to be different from others?
There was a problem hiding this comment.
OK.
Does the number 40000 is a magic number, just need to be different from others?
Yes.
| @@ -193,6 +193,8 @@ var ( | |||
| PerformanceSchemaName = model.NewCIStr("PERFORMANCE_SCHEMA") | |||
| // MetricSchemaName is the `METRICS_SCHEMA` database name. | |||
| MetricSchemaName = model.NewCIStr("METRICS_SCHEMA") | |||
| // SysSchemaName is the `sys` database name. | |||
| SysSchemaName = model.NewCIStr("sys") | |||
There was a problem hiding this comment.
Good question, I didn’t think about it before. I want to know how the information_schema or MySQL 8.0 was introduced at that time. Do you have any advice？ @qw4990
ti-chi-bot
added
needs-rebase
|
@rebelice: PR needs rebase. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## master #22126 +/- ##
================================================
- Coverage 77.0764% 77.0736% -0.0029%
================================================
Files 573 575 +2
Lines 145715 145793 +78
================================================
+ Hits 112312 112368 +56
- Misses 23286 23305 +19
- Partials 10117 10120 +3 |
|
@rebelice: The following tests failed, say
Full PR test history. Your PR dashboard. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here. |
What problem does this PR solve?
Issue Number: #19209
What is changed and how it works?
This is one step of #19209.
Add
sysschema,sys.SCHEMA_UNUSED_INDEXESview andsys.SCHEMA_INDEX_USAGEview.The
sysschema is referencesysschema in MySQL.And
sys.SCHEMA_UNUSED_INDEXESview is also referencesys.SCHEMA_UNUSED_INDEXESin MySQLThe difference between
sys.SCHEMA_INDEX_USAGEandmysql.SCHEMA_INDEX_USAGEis that access tosys.SCHEMA_INDEX_USAGEdoes not need to modify any permissions, and the content ofsys.SCHEMA_INDEX_USAGEhas been managed with permissions, only the index information of the table with theSELECTpermission can be obtained. This is the same forsys.SCHEMA_UNUSED_INDEXES.Related changes
pingcap/docs/pingcap/docs-cn:Check List
Tests
Side effects
Release note
sysschema,sys.SCHEMA_UNUSED_INDEXESview andsys.SCHEMA_INDEX_USAGEview