27 lines
934 B
ReStructuredText
27 lines
934 B
ReStructuredText
|
|
||
|
|
Features
|
||
|
|
========
|
||
|
|
|
||
|
|
The general idea is to provide an audit/versioning trail for important
|
||
|
|
data tables.
|
||
|
|
|
||
|
|
Each table defined in the :term:`app model` can either be versioned,
|
||
|
|
or not. Nothing changes for a non-versioned table.
|
||
|
|
|
||
|
|
For a "versioned" table, a secondary "versions" table is created,
|
||
|
|
schema for which is a superset of the original "versioned" table.
|
||
|
|
When records change in the original table, new "version" records are
|
||
|
|
added to the versions table.
|
||
|
|
|
||
|
|
Therefore you can see how a record has changed over time, by
|
||
|
|
inspecting its corresponding versions.
|
||
|
|
|
||
|
|
When any record changes (for any versioned table), a new "transaction"
|
||
|
|
record is also created. This identifies the user responsible, and
|
||
|
|
timestamp etc. Any new version records will tie back to this
|
||
|
|
transaction record.
|
||
|
|
|
||
|
|
All this is made possible by SQLAlchemy-Continuum; the Wutta-Continuum
|
||
|
|
package mostly just adds config glue. See also
|
||
|
|
:doc:`sqlalchemy-continuum:index`.
|