Indexing & query optimization
Download as PPT, PDF5 likes1,795 views
The document provides an overview of indexing and query optimization in MongoDB. It discusses indexing 101 including why indexes improve performance over table scans, how to analyze query performance using the profiler, and different types of indexes like compound, unique, multikey, covered and geospatial indexes. It also covers creating, listing, dropping indexes and building indexes in the background. The document emphasizes understanding your workload, profiling slow queries, and creating indexes to support slow operations.
![Indexing & Query Optimization Jared Rosoff ( [email_address] )](https://image.slidesharecdn.com/indexingqueryoptimization-110208160038-phpapp02/85/Indexing-query-optimization-1-320.jpg)
![Use explain query = db.coll.find({title:”My blog”}) query.explain(); { "cursor" : " BasicCursor ", "indexBounds" : [ ], "nscanned" : 57594, "nscannedObjects" : 57594, "n" : 3, "millis" : 108 }](https://image.slidesharecdn.com/indexingqueryoptimization-110208160038-phpapp02/85/Indexing-query-optimization-13-320.jpg)
![Embedded documents db.posts.save({ title: “My First blog”, comments: [ {author: “James”, ts : new Date()} ] }); db.posts.ensureIndex({“ comments.author ” : 1})](https://image.slidesharecdn.com/indexingqueryoptimization-110208160038-phpapp02/85/Indexing-query-optimization-18-320.jpg)
![Multikeys { “ tags ” : [ “ mongodb ” , “ cool ” ], ...} db.posts.ensureIndex({ “ tags ” : 1})](https://image.slidesharecdn.com/indexingqueryoptimization-110208160038-phpapp02/85/Indexing-query-optimization-19-320.jpg)
Indexing & query optimization
- 1. Indexing & Query Optimization Jared Rosoff ( [email_address] )
- 2. Overview Indexing 101 Profiling your queries Creating Indexes Managing Index
- 3. Indexing 101
- 4. Ack! My queries are slow! http://michaelprescott.typepad.com/.a/6a00d83451574c69e20134858a87a2970c-800wi
- 5. Index FTW! http://musformation.com/pics/excited.jpg
- 6. Why did that happen? Magic More Magic
- 7. Table scans 1 2 3 4 5 6 7 Looked at 7 objects Find where x equals 7
- 8. Tree Lookup 7 6 5 4 3 2 1 Looked at 3 objects Find where x equals 7
- 9. O(n) vs. O(log n) Table scan Index Number of records Number of comparisons
- 11. Using the Profiler db.setProfilingLevel( level ) 0 == off 1 == slow operations (>100ms) 2 == all operations
- 12. Profiler Output db.system.profile.find ({millis:{$gt:5}}); { "ts" : "Thu Jan 29 2009 15:21:27 GMT-0500 (EST)" , "info" : "query test.foo ntoreturn:0 exception bytes:53" , "millis" : 88 }
- 13. Use explain query = db.coll.find({title:”My blog”}) query.explain(); { "cursor" : " BasicCursor ", "indexBounds" : [ ], "nscanned" : 57594, "nscannedObjects" : 57594, "n" : 3, "millis" : 108 }
- 14. Creating Indexes
- 15. Index a field db.posts.ensureIndex( { ‘name’: 1 }) 1 = ascending -1 = descending
- 16. Compound indexes db.posts.ensureIndex({name: 1, date: -1})
- 17. Unique Indexes db.posts.ensureIndex({title: 1}, {unique: true})
- 18. Embedded documents db.posts.save({ title: “My First blog”, comments: [ {author: “James”, ts : new Date()} ] }); db.posts.ensureIndex({“ comments.author ” : 1})
- 19. Multikeys { “ tags ” : [ “ mongodb ” , “ cool ” ], ...} db.posts.ensureIndex({ “ tags ” : 1})
- 20. Covered Indexes New in 1.7.4 Query can be resolved in index only Need to exclude _id from items projected db.posts.ensureIndex({“title”: 1}) db.posts.find({“title”: “My blog post:}, {title: 1, _id:0}))
- 21. Sparse Indexes
- 23. Managing Indexes
- 26. Background building db.posts.ensureIndex(..., {background: true})
- 27. Query Planning
- 28. Query Planning
- 29. When isn’t an index used?
- 30. Picking an Index scan Index on x Index on y Remember Terminate
- 31. Review Understand your workload Profile your queries Use explain on the slow ones Create indexes for slow operations