This paper presents an approach, the IHF approach, of organising data of a hierarchical structure into one single file, maintaining adequate representation of the relationships among the data. The IHF approach requires each data item be stored in the basic form of name-value pair and consists in assigning each data item with identifier which provides not only the data item identification but also an implicit identification of the hierarchical structural relationships among groups of data items. Data selection and retrieval operations are effectively implemented with the concept of masking of relevant identifier bits, without resorting to any linking devices. Application of the IHF technique on CAFS is examined.