Placing data (cases, for example) from one classification (scheme) into categories of a second classification (when there is one-to-one correspondence) or by concatenating one or more categories from the first classification into the second, a sort of “compound grouping.” Where the groups of the second classification are more specific than or different from those of the first classification, recoding should not be attempted, because no exact recoding is possible; data once coded to a category cannot be “split” without going back to more detailed sources.