Re: [opendx-users] Splitting a field in two

[Torsten Hofmann]

Sharon Gibson wrote:
Seems as though you should just import your data as two fields, since none of
the 6 components appear to be shared. There's probably not any memory savings
in the 1-field approach, and you wouldn't have to do any of the Replace

I'm not trying to save memory.

I'm generating the data in a program of my own and writing it out in 
native opendx format. I thought it's much more convenient to write all 
the results into one 'big' file and then selecting the stuff in opendx 
instead of having to maintain several output files in my external 
program. But it looks like that assumption was wrong.

Or am I overlooking something?

tinkering (which can be tricky when working with attributes, as well as with
invalid positions, etc.)

Yes, the problems arise when dealing with attributes like 'ref' or 
'dep' as you have to set them manually (in a three step process as I 
described earlier, see below).

Torsten

On Dec 17,  4:48pm, Torsten Hofmann wrote:

Subject: Re: [opendx-users] Splitting a field in two
Donna,

thank you very much for your answer.

Sadly what you suggest doesn't work - or maybe you misunderstood what
I wanted to achieve :-)

field=Replace(field, field, "other-positions",   "positions")
field=Replace(field, field, "other-connections", "connections")
field=Replace(field, field, "other-data",        "data")
field=Remove(field, {"other-positions", "other-connections", "other-data")

is equivalent to

field=Rename(field, "other-positions",   "positions")
field=Rename(field, "other-connections", "connections")
field=Rename(field, "other-data",        "data")

In both cases the problem remains that the attributes are preserved.
So the new 'connections' component has an attribute 'ref' with the
value 'other-positions' and not 'positions'. The new 'data' component
has an attribute 'dep' with the value 'other-positions' which should
now be 'positions'.

In order to fix this, I have to

data=Extract(field,"data")
data=Options(data,"dep","positions")
field=Replace(field,data,"data","data")

and the same for the connection component.

So again, is there a shorter/easier way to achieve the overall result?

Torsten

Donna L Gresh wrote:

You might try Replace instead of Rename. I think the syntax is
 Replace(your-6-field, your-6-field, "positions", "other-positions");
//or maybe the last two switched

this should preserve the attributes (I think :)



Donna L. Gresh, Ph.D.
Optimization and Mathematical Software Group
IBM T.J. Watson Research Center
(914) 945-2472
http://www.research.ibm.com/people/g/donnagresh
[EMAIL PROTECTED]



        *Torsten Hofmann <[EMAIL PROTECTED]>*
Sent by: [EMAIL PROTECTED]

12/16/2003 10:48 AM
Please respond to opendx-users


       To:        opendx2-users@lists.berlios.de
       cc:
       Subject:        [opendx-users] Splitting a field in two




Hi,

I have a field with 6 components:
-positions
-connections
-data
-other_positions
-other_connections
-other_data

where the first and last three components each belong together. I'd
like to visualize the two datasets separately, basically splitting
this into two fields with 3 components each.

What I do at the moment to visualise the 'other_' dataset is: Rename
other_positions to positions, other_connections to connections and
other_data to data. The problem is that afterwards I still have to fix
the attributes 'ref' of connections and 'dep' of data to point to the
correct component and not to the 'other_' component. This takes each
an Extract, Options, Replace cycle.

This works but looks and feels a bit convoluted to me. Is there an
easier way?

Thanks a lot for your help.

Torsten

--
Dipl. Rech. Wiss. ETH Torsten Hofmann
Lehrstuhl Werkstoffkunde und Technologie der Metalle (WTM)
Universität Erlangen-Nürnberg
Martensstr. 5
91058 Erlangen

Phone : +49 9131 / 85 287 26
Fax   : +49 9131 / 85 275 15
Email : [EMAIL PROTECTED]

-- End of excerpt from Torsten Hofmann