Using full (cartesian) joins could easily explode the volume of matching taking place. To limit this a range of composite keys can be created based on various combinations of the matching fields.
A full composite may include the full name and postcode (JOHNSMITH20001225NE158NY) and would only join on corresponding records with the same composite (e.g. a full match).
A composite based on initials and postcode area (JSNE) would join to everyone with a matching composite in the other dataset.
The composites should complement the various potential confident match criteria. For example, it is possible to have a confident match whilst missing a postcode so we need to make sure the postcode is not include in all composite keys.
Using full (cartesian) joins could easily explode the volume of matching taking place. To limit this a range of composite keys can be created based on various combinations of the matching fields.
A full composite may include the full name and postcode (JOHNSMITH20001225NE158NY) and would only join on corresponding records with the same composite (e.g. a full match).
A composite based on initials and postcode area (JSNE) would join to everyone with a matching composite in the other dataset.
The composites should complement the various potential confident match criteria. For example, it is possible to have a confident match whilst missing a postcode so we need to make sure the postcode is not include in all composite keys.