Failing Associations

TCA obtained a perfect score because the data was so perfectly regular. Real life is not that tidy. For example, what happens if no association is found for a particular primary? This can be done simply by shutting down the time gate using the commands:-

$tca_time_gate -1. 1.
@tca_demo

Now the output will contain:-

       Output: File: tca_demo_out.ntp
               N-tuple ID:    201
               No. of entries written:           0

showing that all associations failed. However, suppose you want to study the failures to see if they have some common feature, then you want TCA to output something even for failures. TCA can do this using the concept of the undefined association. In this case any variable in the output n-tuple that is derived, directly or indirectly, from the secondary, will be set to the undefined value, but any variable coming from the primary will be defined. This can be demonstrated with the commands:-

$tca_output_unass $on
$tca_time_gate -1. 1.
@tca_demo

and now the output shows:-

       Output: File: tca_demo_out.ntp
               N-tuple ID:    201
               No. of entries written:          10

but the next two lines are:-

               Of which:-
                         10    Unassociated primaries

which shows that they are all unassociated.