3. Writing collections JEVEUX#

3.1. Contiguous collections#

Contiguous collections are built from simple objects, so it’s easy to use the methods associated with simple objects to store them. Contiguous collection objects are stored in the value segment associated with the system object $$ DESO.

The simple collection descriptor object is stored in a « dataset » with the name of the collection under the « group » /.

A list named “ATTRIBUTS JEVEUX” of 5 attributes is associated with the associated « dataset » containing respectively:

  1. a text (“COLLECTION”),

  2. an identifier (simple object identifier JEVEUX),

  3. a string containing the class, genre (X), and type in the JEVEUX (I) sense,

  4. the associated Fortan type,

  5. an empty string.

Collection system objects are stored in « datasets » with the name of each collection system object under the « group » /.

A list named “ATTRIBUTS JEVEUX” of 5 attributes is associated with the associated « dataset » containing respectively:

  1. a text (“OBJ. SYSTEME COLLECTION “),

  2. an identifier (simple object identifier JEVEUX),

  3. a string containing the class, genre, and type in the JEVEUX sense,

  4. the associated Fortan type,

  5. an empty string.

3.2. Scattered collections#

Dispersed collections are built from simple objects for the collection system objects and value segments associated with each collection object. System objects are stored in the same way as for contiguous collections. The system object $$ DESO is particular because it is not associated in the case of scattered collections with a segment of values, so there is no associated « dataset », and it is only by loading the content of the system objects JEVEUX that the scattered collection is rebuilt when rereading the file HDF.

A « group » with the name of the collection supplemented by __ OBJETS is created at the level of the « group »/to write scattered collection objects. A list named “ATTRIBUTS JEVEUX” of 5 attributes is associated with the « group », only the first element is non-empty and contains text (COLLECTION).

Each collection object is then written in a "dataset" with the name of the collection supplemented by the number of the collection object (including for named collections) under the "group" described above.

A list named “ATTRIBUTS JEVEUX” of 5 attributes is associated with the associated « dataset » containing respectively:

  1. a text (OBJET FROM COLLECTION),

  2. the name or number of the collection object and the collection identifier,

  3. a string containing the class, genre, and type in the JEVEUX sense,

  4. the associated Fortan type,

  5. an empty string.