Thank you for the reply. So, I want to combine a collection of .dag files into a single DAG and then add a few children post-processing nodes. I have written the following toplevel DAG which accomplices this:
# begin combined.dag
SPLICE H1 pipeline-1.dag DIR ../pipeline-1
SPLICE H2 pipeline-2.dag DIR ../pipeline-2SPLICE H3 pipeline-3.dag DIR ../pipeline-3
JOB PPROC1 PPROC1.sub DIR ../pproc1
VARS PPROC1 ...
PARENT H1 H2 H3 CHILD PPROC1
JOB PPROC2 PPROC2.sub DIR ../pproc2
VARS PPROC2 ...
PARENT PPROC1 CHILD PPROC2
The number of pipeline .dag's and the number of post-processing nodes (as well as the values of their variables) is not fixed, so I want to be able to write this dag file programmatically, and I am trying to achieve the above result using the HTCondor Python Bindings. I believe I could use a SUBDAG instead of a SPLICE, but reading the HTCondor manual I got the impression that I should try to use a SPLICE whenever possible.