Underkilling is a data cleaning method that involves removing duplicates between multiple files (or within the same file), while potentially leaving some duplicates. This deduplication approach is characterized by the fact that it operates only on fields with identical content, which can lead to some data being retained. For example, an individual named “Stephan” and another named “Stephane” sharing the same name and address will not be merged and will both remain in the file.

Sometimes, underkilling is deliberately performed to avoid excessive data deletion, for fear of losing addresses during a data cleaning operation. This approach allows for a certain margin of error and preserves potentially useful data, even if it may result in duplicates remaining in the files.