20ranging

20ranging is a conceptual data indexing technique intended to speed up multi-attribute range queries on large data collections. The term originates from partitioning each numeric attribute into a fixed number of equal-width intervals, typically twenty, creating a coarse representation of the attribute space called twenty-ranges. This representation supports fast filtering, with exact results obtained through subsequent refinement on candidate items.

Operation: Each attribute is divided into 20 bins. A data record is labeled by its bin indices across attributes, producing a multi-bin signature. A query with ranges across attributes is converted into a set of bin combinations that may contain matches. The system retrieves records from those bins and then verifies them precisely. Additional structures such as histograms or compact filters help prune unlikely bins.

Applications and variants: 20ranging is discussed in contexts such as data warehousing, exploratory analytics, and approximate query processing. Variants include dynamic bin boundaries to adapt to data distribution and integration with traditional indexes to support exact lookups after pruning.

Limitations: Its effectiveness depends on data distribution and query shapes. Coarse binning can lead to many candidates in dense regions, and bins may need reconfiguration as data evolves. The approach trades exactness for speed and is most suitable when range predicates dominate queries and approximate pruning suffices for initial filtering.