On Oct 14, 12:23 am, "Bill Renaud"
wrote:
Aha, sounds like another electrical engineer at work! (Me too, 20 years
ago!) I once did this very thing to save us time checking the accuracy of a
PC (printed circuit) layout output file from REDAC software (I used Pascal
to write it). (I notice that your output has an error in it: Group 3 has
connection 45 in it, which should make it a member of group 2!)

I have 2 suggestions for the output format before working on this
algorithm:

1. Convert your list into 2 columns (this might be the easiest):

Group Connection
----- ----------
1 2
1 3
1 9
2 7
2 8
2 12
2 30
2 45
3 56
3 65

2. Convert your list into 2 columns, but condense column 2 into a single
cell (probably easier for an end user to read; apply word-wrap to the final
worksheet formatting):

Group Connections
----- --------------------------------
1 2,3,9
2 7,8,12,30,45
3 56,65

Which format would be preferable?
--
Regards,
Bill Renaud

:) You're right in saying that I'm an electrical engineer... You're
really experienced in reading people's minds I guess... Thank you for
the suggestion... Actually I'm sorry about that Group 2 issue... I did
a mistake while typing...

Actually the first option looks good though I still want to preserve
the parent-child relation here. For example, from the above list, I
want to retain the least information that when 2,3 and 3,9 belong to
the group, 2 and 3 are directly related; 3 and 9 are directly related
instead of falling into the assumption that 2 and 9 are directly
related. But either ways its fine... Would you suggest a workflow for
this please?