If only the index name, the same name that the situation will lead to more efficient queries reduce; you might say, to build a age of le index does not have to, of course not, because mongo a query can only use one index, so like this situation needs to establish a composite index to improve query efficiency.

To delete that why there is identical to the first key-key index it, it comes on top of a query can only use one index, and the composite index may use a prefix query instead of a single name index, so the index is a single name a waste.

3, multi-key indexing: indexing the array.

语法：db.collectionName.createIndex({'address.city':1})

4, hash indexes:

语法：db.collectionName.createIndex({'name':'hashed'})

Hash index on the inlet are uniformly distributed, useful slice set.

Index attributes:

Create an index:

db.collectionName.createIndex(
    {'name':1},
    {
        'background':true,
        'unique':true,
    'name':'indexName',
    'sparse':true
    }
)

Delete the index:

1, delete the specified name: db.collectionName.dropIndex ( 'indexName')

2, delete indexes on the set (_id cut can not): db.collectionName.dropIndexs ()

3, rebuild the index on the set: db.collectionName.reIndex ()

4, the index set of queries: db.collectionName.getIndexs ()

MongoDB query optimization:

1, slow queries positioning

Open slow query: db.setProfilingLevel (n, {m})

n has three possible values:

0; not recorded (the default value).
1; record slow query log, it must be specified as 1 m, that is, the query time threshold, unit ms.
2; log all slow query log.

After the slow query MongoDB open, the log will be stored in system.profile of the collection, its maximum allocation 128K space.

May be used to sort $ nartual slow query log, such as: db.system.profile.find () Sort ({ 'Natural $': -. 1}).. Limit (. 5)

2, slow query analysis

Can be analyzed by the query plan explain slow query, such as db.collectionName.find ({ 'age': { '$ lt': 50}}). Explain ( 'executionStats')

explain optional parameters:

queryPlanner; the default, meaning only show the execution plan information.
executionStats; represents information showing the execution plan to showcase the implementation of information selected execution plan.
allPlansExecution; represents information showing the execution plan, and present information on the implementation of the selected execution plan, the implementation of information also shows an alternative execution plan.

3, using the index on recommendations

According to demand the establishment of the index, it will be useful but there are costs, not to those who write once read many less indexing .

Try to ensure that each stage of the query are IXSCAN , the pursuit of the number of scanning a document (totalDocsExamined) = Returns the number of (nReturned) document .

MongoDB in a query with only an index, if the multi-criteria query to make use of the composite index .

Indexing the amount of data is very time consuming and more resources, so try to put a small amount of data indexing built in the time .

MongoDB study notes (five, MongoDB storage engine and index)

table of Contents:

mongoDB storage engine:

mongoDB Index:

Index attributes:

MongoDB query optimization:

Guess you like