Why Isn't My Query Using an Index?: An Introduction to SQL Performance
0 likes395 views
The document discusses SQL performance with a focus on why Oracle may not use indexes for queries. It provides insights into full table scans versus index range scans, and highlights the importance of clustering factors for efficient data access. The author, Chris Saxon, emphasizes the trade-offs involved in different access methods and offers guidance on improving query performance.
Why Isn't My Query Using an Index?: An Introduction to SQL Performance
- 1. Copyright © 2015 Oracle and/or its affiliates. All rights reserved. | 1 Why Isn't My Query Using an Index? An introduction to SQL performance Chris Saxon, Oracle Developer Advocate for SQL [email protected] @chrisrsaxon & @sqldaily blogs.oracle.com/sql www.youtube.com/c/TheMagicofSQL
- 2. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | Why isn’t Oracle using my #@!% index?!
- 3. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
- 4. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
- 5. Colour BAG Colour BAG Colour BAG Colour BAG Colour BAG Blue 73 Brown 56 Green 11 Orange 19 Red 2 Blue 31 Brown 40 Green 31 Orange 19 Red 41 Blue 1 Brown 41 Green 70 Orange 85 Red 21 Blue 64 Brown 65 Green 4 Red 42 Red 32 Blue 69 Brown 23 Green 80 Red 100 Red 12 Blue 85 Brown 21 Green 28 Red 70 Red 83 Blue 84 Brown 73 Green 44 Red 38 Yellow 29 Blue 40 Brown 97 Green 99 Red 33 Yellow 31 Blue 41 Brown 57 Green 32 Red 63 Yellow 51 Blue 88 Brown 63 Orange 85 Red 22 Yellow 33 Blue 100 Brown 100 Orange 21 Red 80 Yellow 4 Blue 6 Brown 80 Orange 76 Red 80 Yellow 9 Blue 64 Brown 75 Orange 32 Red 40 Yellow 27 Blue 90 Green 20 Orange 93 Red 68 Yellow 48 Blue 24 Green 50 Orange 70 Red 54 Yellow 56 Brown 88 Green 85 Orange 2 Red 93 Yellow 15 Brown 50 Green 17 Orange 15 Red 93 Yellow 81 Brown 69 Green 27 Orange 5 Red 20 Yellow 83
- 6. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | What’s this got to do with ?
- 7. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | M&M => a row Document => an index A bag => a block
- 8. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | Method #1 => TABLE ACCESS FULL Method #2 => INDEX RANGE SCAN Multiblock reads Single block reads
- 9. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | Source: Pixabay Which is faster?
- 10. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | FULL TABLE SCAN Visit each bag once INDEX RANGE SCAN Visit each bag ~10 times
- 11. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | What if there’s a Promotion? silver sweets in 1% of bags
- 12. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | FULL TABLE SCAN Visit each bag once INDEX RANGE SCAN One bag visit
- 13. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | The Cost Based Optimizer estimates these calculations
- 14. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | DEMO
- 15. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | "M & Ms (3685102458)" by Greg Willis from Denver, CO, usa Uploaded by russavia. CC BY-SA 2.0
- 16. Colour BAG Colour BAG Colour BAG Colour BAG Colour BAG Blue 1 Brown 7 Green 13 Orange 19 Red 25 Blue 1 Brown 7 Green 13 Orange 19 Red 25 Blue 1 Brown 7 Green 13 Orange 19 Red 25 Blue 2 Brown 8 Green 14 Red 20 Red 26 Blue 2 Brown 8 Green 14 Red 20 Red 26 Blue 2 Brown 8 Green 14 Red 20 Red 26 Blue 3 Brown 9 Green 15 Red 21 Yellow 27 Blue 3 Brown 9 Green 15 Red 21 Yellow 27 Blue 3 Brown 9 Green 15 Red 21 Yellow 27 Blue 4 Brown 10 Orange 16 Red 22 Yellow 28 Blue 4 Brown 10 Orange 16 Red 22 Yellow 28 Blue 4 Brown 10 Orange 16 Red 22 Yellow 28 Blue 5 Brown 10 Orange 17 Red 23 Yellow 29 Blue 5 Green 11 Orange 17 Red 23 Yellow 29 Blue 5 Green 11 Orange 17 Red 23 Yellow 29 Brown 6 Green 11 Orange 18 Red 24 Yellow 30 Brown 6 Green 12 Orange 18 Red 24 Yellow 30 Brown 6 Green 12 Orange 18 Red 24 Yellow 30
- 17. Colour BAG Colour BAG Colour BAG Colour BAG Colour BAG Blue 1 Brown 3 Green 4 Orange 7 Red 9 Blue 1 Brown 4 Green 4 Orange 8 Red 9 Blue 2 Brown 5 Green 5 Orange 9 Red 9 Blue 2 Brown 5 Green 6 Red 1 Red 9 Blue 3 Brown 5 Green 6 Red 2 Red 10 Blue 4 Brown 6 Green 7 Red 2 Red 10 Blue 6 Brown 7 Green 8 Red 3 Yellow 1 Blue 7 Brown 8 Green 8 Red 3 Yellow 1 Blue 7 Brown 8 Green 10 Red 4 Yellow 2 Blue 8 Brown 9 Orange 1 Red 4 Yellow 2 Blue 9 Brown 9 Orange 2 Red 5 Yellow 3 Blue 9 Brown 10 Orange 3 Red 6 Yellow 5 Blue 10 Brown 10 Orange 3 Red 6 Yellow 5 Blue 10 Green 1 Orange 4 Red 7 Yellow 6 Blue 10 Green 2 Orange 4 Red 7 Yellow 7 Brown 2 Green 3 Orange 5 Red 8 Yellow 8 Brown 2 Green 3 Orange 6 Red 8 Yellow 9 Brown 2 Green 3 Orange 6 Red 8 Yellow 10
- 19. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | clustering factor = 10 (good)
- 20. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | clustering factor = 120 (bad)
- 21. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | DEMO
- 22. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | How do I Improve the clustering factor?
- 23. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | Tannin Heap Simon A. Eugster CC-BY-3.0 By default tables are
- 24. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | Organized Part ition ed Clusters How do I ensure consistent clustering?
- 25. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | There can be only 1!
- 26. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | Colour Weight
- 27. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | Cost Rows Full Table Scan Index Range Scan 1 Index Range Scan 2 Index Range Scan 3
- 28. Rows “Toy Blocks (6831034024)” welcometolearn CC BY 2.0
- 29. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | developer.oracle.com #MakeDataGreatAgain