1. Brief introduction
The index is to speed up the query. Given that there is no index, MONGODB does a table scan when querying, and if the collection is very large, this query is very slow. The keys are generally indexed when the query is created.
Index the Sort field, assuming that the Sort,mongodb field will take all the data into memory to sort, assuming that the collection is too large to be sorted in memory, MongoDB will give an error.
2. MongoDB CREATE index
Create an index using the Ensureindex command.
> Db.people.ensureIndex ({"username": 1});
The above statement indexes the username key of the People collection.
3. Combined Index
For composite queries or sorting, set up a composite index.
> Db.people.ensureIndex ({"Date":-1, "username": 1});
The index key is 1 or-1, which indicates the order in which the index is created, 1 is ascending, and 1 is reversed. Assuming that the index has only one key, the direction does not matter.
4. Indexing of embedded documents
The comments for the blog collection are indexed according to time:
> Db.blog.ensureIndex ({"Comments.date": 1});
5. Unique index
A unique index ensures that each key in the collection is a unique value.
> Db.people.ensureIndex ({"username": 1}, {"Uniqe": true});
When you create a unique index on an existing collection, it is possible that you have repeated values, which creates an index failure. Dropdups retains the first document, deleting the values that are repeated later.
> Db.people.ensureIndex ({"username": 1}, {"Unique": true, "dropdups": true});
6. View the established index
The index is placed in the System.indexes collection.
> Db.system.indexes.find ();
The ability to view the index key, name, and which collection belongs to.
7. Deleting an index
With the name of the index, the index can be dropped using dropindexes. The index name is queried by System.indexes.
> db.user.dropIndexes ({"Username_1": 1});
Address: http://blog.csdn.net/yonggang7/article/details/28100855