Semantic integrity refers to the extent to which different aspects of a data record or other representation makes sense internally, This may refer to an individual field or unit in the data for example, if a record states that the age of a schoolchild is 73, or the relationship between items, for example if the birth date of a historical person is recorded as after their death.
During data cleaning one can perform various semantic integrity checks alongside sanity checks on the original data, triggering further processing or human scrutiny. Semantic integrity checks can also be used after data processing steps to ensure they have worked correctly.
Used in Chap. 10: page 133