In other words, tables which are only Parents of other tables and which are not Children of any others will never cause any rows to be extracted because they will never have any rows themselves.

The above situation is undesirable because it can leave large numbers of tables unpopulated by the Extraction Set rules. The Reverse Tails Rules generator suggests rules that are the exact reverse of the existing Parent-Child relationship. When combined with the existing rule a Reverse Tails rule means the Parent requires rows to be present in the Child and the Child requires rows to be present in the Parent. Sometimes this is a good thing and sometimes (depending on the join key) this can lead to a race condition in which the extraction of the entire contents of both tables is the only way of satisfying the rule pair. Reverse Tails rules should be added to the Extraction Set one or two at a time to populate tables that are omitted by the current Extraction Set rules.

The DataBee Reverse Tails Dialog Box

The Reverse Tails Rules dialog box is very easy to use. The DataBee Set Designer application calculates all of the candidate rules as the dialog box starts. All that remains is to select the rules you wish to add to the set and then to press the Add Selected Rules to Set button.

The Src. Rows and Ext. Rows columns show the number of rows in the source schema and the number of rows in the last extract for the Parent table of the proposed new rule. The number of rows in the last extract can be used as a guide to how well the reverse tails rule will populate the target table.

Some Notes on using the Reverse Tails Rules

1. Never add Reverse Tails rules at random - add them in for a reason.
2. Never add all Reverse Tails rules at once - it is better to add one or two and observe the effect.
3. Always perform a test extract with the Extraction Set after adding Reverse Tails rules to see how the newly added rules have affected the table row counts.
4. Reverse Tails Rules are always added in with rule numbers in the range 3000 to 3499 so it is quite easy to identify rules (in the Rules Tab) that were created via this method.
5. Another way to populate a table that is the Parent of other tables but the Child of none is to define a WHERE Clause rule for the table.