1.3.4.9.22 Match Transformation: Soundex

The Soundex transformation creates common soundex keys from values that sound the same, but may be different, for example, due to misspellings.

The Soundex transformation is similar to the Metaphone transformation, but uses a different way of detecting whether two values sound the same. It is generally 'looser' in terms of how similar two values have to sound in order for the same key to be generated (for example, generating the same key for "Smith" and "Snaith", which have different metaphone keys.

Also, and importantly, the Soundex transformation only operates on single words (or, when processing multi-word values, on the first word). This means that "Margaret Hawkins" and "Margaret Johnson" would generate the same soundex key (M626), but different metaphone keys.

The Soundex transformation is useful when clustering or matching against a single word name identifier, such as either First Name or Surname. Use it where spelling mistakes are common in key identifiers such as names, and you need to ensure that any names that could be the same are caught by the matching rules.

Options

None

Example

In this example, the Soundex transformation is used to create the initial clusters for a small dataset by transforming the values for a Surname identifier into their soundex keys.

Example transformations

The following table shows example Soundex transformations:

Table 1-95 Example Transformations for Soundex

Value	Transformed Value
Howard	H630
Hayward	H630
Hardy	H630
Price	P620
Pierce	P620
Preece	P620
Pryke	P620
Roberts	R163
Robertson	R163