Providing a subset of de-duplicated data as output is disclosed. The subset of data includes a plurality of data segments. In some embodiments, the output includes segments of data stored across a plurality of containers in a de-duplicated storage system. Each container stores one or more data segments from the subset as well as one or more segments not from the subset. For each container that includes at least one data segment from the subset, the corresponding container data is included in the output in its entirety, including segments that are not from the subset. In the event that container data is included in the output, a value in a data structure that corresponds to the container is updated. The data structure includes a value for each container in the de-duplicated storage system that indicates whether the corresponding container data has been included in the output.