Equality-Based Indexing

The typical record consists of several attributes: names, addresses, transaction dates, prices, sizes, colors, etc. We expect duplicates to be similar across most attributes. For some attributes, we even expect an exact match—for example, duplicate customer records will unlikely have different country attributes.

Note: The restaurants dataset we use below is open data. See the Glossary of the course for attribution and references.

Standard blocking (SB)

SB, or “blocking,” is so prevalent that indexing is often used as a synonym for this technique. If not stated otherwise explicitly, people mean SB when they talk about indexing on a particular attribute.

Below, we read the restaurants dataset and use recordlinkage to block by the city attribute: