Databases Reference
In-Depth Information
period. For example, a tariff discount has values which
depend on the year quarter. The use of versions in the data
repository for this requirement would not be appropriate as
we are interested in one piece of data;
-
the creation of a version is not necessarily linked to
the notion of time. For example, the creation of a working
branch to test a new configuration of the data repository is
not linked to the time management.
In the same way, the use of data updates by use context is
not suitable because it is necessary to create as many
contexts as time periods. This would make no sense and
would bastardize the normal usage of context which is more
oriented to taking into account geographical and
organizational structures.
Neither must the management of time be confused with
the audit trail function. This tracks, in a log, the different
values which the data hold over time, associating with them
information such as the user or the system from which the
update originated. The audit trail records all updates and
read accesses of all data in the order in which they take
place in the database. There is no specific audit trail for each
data element.
Having separated the ambiguities between versions,
contexts and audit trails, we can now detail the principles of
time management, firstly via the function of data history
tracking and then the management of the business
transactions, and finally the periods of validity.
6.5.1.
Data history tracking
Data history tracking consists of saving, systematically,
all the values assumed by a particular data over time. It is a
kind of audit trail but applied to the scope of only one data
element. For example, an insurance company must save all
