[teacher Zhao Yuqiang] index in mongodb (Part 2)

（ Four ） Index type 3 ： Composite index （Compound Index）

MongoDB Support composite index , To combine multiple keys to create an index . This approach is called composite index , Or composite index , This method can satisfy the situation of using index in multi key value matching query . Secondly, when using composite index , Index can also be used by prefix method .MongoDB The composite index in is basically consistent with the relational database . Some of the principles of composite indexing in relational databases also apply to MongoDB.

In the previous content , We've been emp A composite index is created on the collection , as follows ：

db.emp.createIndex({"deptno":1,"sal":-1})

The following uses different filter conditions to query documents , Check the corresponding execution plan ：

（1） Use only deptno As a filter

db.emp.find({"deptno":10}).explain()

（2） Use deptno、sal As a filter

db.emp.find({"deptno":10,"sal":3000}).explain()

（3） Use deptno、sal As a filter , But put sal On the front

db.emp.find({"sal":3000,"deptno":10}).explain()

（4） Use only sal As a filter

db.emp.find({"sal":3000}).explain()

（ 5、 ... and ） Composite index and sort

Composite indexes are created in ascending or descending order to specify how they are arranged . For single key indexes , The order is not particularly important , because MongoDB You can traverse the index in any direction . For composite indexes , How to sort can determine whether the index can be used in the query .

db.emp.createIndex({"deptno":1,"sal":-1})

In the previous content , We've been deptno In ascending order 、sal A composite index is created in descending order on the , Next, test the order of the next , Whether the index has been executed ：

 When the index is used ：
db.emp.find().sort({"deptno":1,"sal":-1}).explain()
db.emp.find().sort({"deptno":-1,"sal":1}).explain()

 Without using the index ：
db.emp.find().sort({"deptno":1,"sal":1}).explain()
db.emp.find().sort({"deptno":-1,"sal":-1}).explain()

 Swap the positions of two columns , Retest .

（ 6、 ... and ） Composite index and index prefix

The index prefix refers to a subset of a composite index , If there is an index like this ：

db.emp.createIndex({"deptno":1,"sal":-1,"job":1})

 Then there is the following index prefix ：
{"deptno":1}
{"deptno":1,"sal":-1}

stay MongoDB in , In the case of the following query filter conditions , The index will be used to :

db.emp.find().sort({deptno:1,sal:-1,job:1}).explain()
db.emp.find().sort({deptno:1,sal:-1}).explain()
db.emp.find().sort({deptno:1}).explain()

In the case of the following query filter conditions , The index will not be used to :

db.emp.find().sort({deptno:1,job:1}).explain()
db.emp.find().sort({sal:-1,job:1}).explain()

（ 7、 ... and ） Summary

• Composite indexes are based on multiple keys ( Column ) Index created on
• Composite indexes can be created for each key ( Column ) To specify the sort method
• The sorting method of the key column affects the operation of the query when sorting , Only those in the same or opposite direction can be matched
• Composite index and prefix index can only be used if they match