Model
How wide a block's min–max range is
A zone map is not an index. It is a measurement of how well you sorted.
- Block size
- rows per block, values uniform on [0, 1]
How it behaves
One input on the horizontal axis, the others held fixed.
Data table
| Rows per block b | Range covered |
|---|---|
| 1 | 0% |
| 1.19 | 8.61% |
| 1.41 | 17.1% |
| 1.68 | 25.3% |
| 2 | 33.2% |
| 2.37 | 40.7% |
| 2.82 | 47.6% |
| 3.35 | 54% |
| 3.98 | 59.8% |
| 4.73 | 65.1% |
| 5.62 | 69.8% |
| 6.68 | 74% |
| 7.94 | 77.6% |
| 9.44 | 80.8% |
| 11.2 | 83.6% |
| 13.3 | 86% |
| 15.8 | 88.1% |
| 18.8 | 89.9% |
| 22.4 | 91.4% |
| 26.6 | 92.8% |
| 31.6 | 93.9% |
| 37.6 | 94.8% |
| 44.7 | 95.6% |
| 53.1 | 96.3% |
| 63.1 | 96.9% |
| 75 | 97.4% |
| 89.1 | 97.8% |
| 106 | 98.1% |
| 126 | 98.4% |
| 150 | 98.7% |
| 178 | 98.9% |
| 211 | 99.1% |
| 251 | 99.2% |
| 299 | 99.3% |
| 355 | 99.4% |
| 422 | 99.5% |
| 501 | 99.6% |
| 596 | 99.7% |
| 708 | 99.7% |
| 841 | 99.8% |
| 1K | 99.8% |
| 1.19K | 99.8% |
| 1.41K | 99.9% |
| 1.68K | 99.9% |
| 2K | 99.9% |
| 2.37K | 99.9% |
| 2.82K | 99.9% |
| 3.35K | 99.9% |
| 3.98K | 99.9% |
| 4.73K | 100% |
| 5.62K | 100% |
| 6.68K | 100% |
| 7.94K | 100% |
| 9.44K | 100% |
| 11.2K | 100% |
| 13.3K | 100% |
| 15.8K | 100% |
| 18.8K | 100% |
| 22.4K | 100% |
| 26.6K | 100% |
| 31.6K | 100% |
| 37.6K | 100% |
| 44.7K | 100% |
| 53.1K | 100% |
| 63.1K | 100% |
| 75K | 100% |
| 89.1K | 100% |
| 106K | 100% |
| 126K | 100% |
| 150K | 100% |
| 178K | 100% |
| 211K | 100% |
| 251K | 100% |
| 299K | 100% |
| 355K | 100% |
| 422K | 100% |
| 501K | 100% |
| 596K | 100% |
| 708K | 100% |
| 841K | 100% |
| 1M | 100% |
Worked examples, recomputed
Every number the mechanisms quote for this relationship is recomputed from it, so none of them disagrees with it.
Where it appears
Each mechanism states it in its own names. The general symbol each name stands for is shown with it.
- Indexing
When does a separate "where is it?" structure pay for itself, and what does it cost to keep it true?
- Query Planning & Optimization
How does a system choose a plan for a question when the facts that decide the best plan are unknown until the plan has run?
B (rows per block) =