Next: Data Repositories
Up: Hierarchical Data Storage
Previous: FITS Files and Data
Contents
The main objective of the usage of Data Objects is an easy and efficient access to the data, you
should just specify the name of the desired Data Structure. However, there is an ambiguity as soon
as two Data Elements have the same name. For these reasons the concept of Index was introduced.
An Index is a FITS Group that, in addition to the usual stuff stored in FITS Groups, may contain
columns describing the peculiarities of the Children. All Children in the Index Group have the
same name. In ISDC Data Model only two cases are possible:
- all direct members of the group have different names;
- all direct members of the group have the same names. The group is therefore an Index.
In case 1, we may have to store several Data Elements that have the same name in the Object.
However, and that is the reason it is stated ``direct members'', these Data Elements must be
encapsulated in an Index.
Standard Groups are those whose direct members all have different names. Apart from the fact
that direct members must all have different names, the only constraint is that the group name
should follow the pattern:
XXXX-XXXX-GRP
All members of the index table are identical, but contain the additional information that helps to
identify them. The name of Index Group follows the pattern:
XXXX-XXXX-XXX-IDX,
where XXXX-XXXX-XXX is the name of the children data structures.
Among other things, an Observation Group contains an Index, which members are Science Window
Groups. An Observation Group must always contain this Index Group whose members are
identical. This index allows the user to search effciently for Science Window Groups using the
Science Window ID, time, and so on. The structure of an Observation Group is illustrated in
Figure 10.
Next: Data Repositories
Up: Hierarchical Data Storage
Previous: FITS Files and Data
Contents
inthelp@sciops.esa.int