The Key Generator transformation uses a grouping strategy to create group keys for the column you select. The strategies are String, Soundex, and NYSIIS. Records with common values in the selected field have a common group key value. The Match transformation processes records with common group key values together. This enables faster duplicate analysis in the Match transformation.
The number of comparison operations that the Match transformation must perform grows exponentially with the number of records in the data set. This exponential growth can consume significant amounts of computing resources. By creating group keys, the Key Generator transformation enables the Match transformation to compare records in smaller groups, which reduces processing time.
When you perform field matching, select a column for group key generation that is likely to provide useful groups for your matching needs. For example, a Surname column is likely to provide more meaningful group key data than a Forename column. But do not use the Surname column if you intend to select that column for duplicate analysis in the Match transformation.
The Key Generator transformation can also create a unique ID for each record. Each record that enters the Match transformation must contain a unique ID. Use the Key Generator transformation to create IDs for your data if none exist.