Dear All

I am trying to extract data from several groups and merge to selected data into a single group. The data file structure looks like:

member0            (MainGroup)
   member0         (SubGroup - FEdata)
       positions   (DX field - positions, connections, data...)
       connections
       data etc.
   member1
       positions
       connections
       data etc.
   etc.
member1
   member0
       positions
       connections
       data etc.
   member1
       positions
       connections
       data etc.
   etc.
member3
  member0
etc.

What I need to do is select, for example, member 0 and member 3 (MainGroups 01 & 04 and have access via a single <select>/<selector> pair to all the members within the 2 groups from a single <select>/<selector> pair.

If it helps, the highest level members (MainGroups) are groups of finite elements, and the second level members (SubGroups) contain the various data results (displacement, stress component etc.) for each set of elements.

Thus, I'd like to present one <selectorlist> that says "select the regions of interest" and another that asks "which parameter would you like to investigate". Unfortunately, I haven't worked out how to merge all the groups data from several top level members into a single group.

I've tried using 2 macros with a <foreachmember> loop in the first, calling a second one with a <forN> which iterates each second level member. However the <setlocal>/<getlocal> pairs coupled to an <append> in the first macro recombine the data into the original form, even though the within the macro everything is separated as I want them to be

The results are written from a Perl script, so if changing the structure of DX file there is a better way to go than doing it inside DX that is possible. Unfortunately, there is a limit to mow many field can be displayed in a <selectorlist> so I cannot just write all my FE data fields (upto 20) for all my element groups (often 20+) into a flat file. Since that exceeds the limit (and makes for really huge data sets to visualize).

Thanks for your time

Peter

Reply via email to