Files represent sequences of bytes. Unfortunately, accessing the bytes is not generally sufficient for most file processing. It is typically necessary to know how to interpret the bytes as meaningful pieces of data. In DataFlow this information is provided by formats and schemas.