Dear all
I have a 3D "spreadsheet" style file (imported via the *.general
filter) in which one column contains a string identifier:
i.e.:
X , Y, Z, data1, data2, data3, ID
1 1 1 25 26 27 set1
6 30 7 27 24 29 set2
1 2 7 26 25 28 set1
4 20 3 26 25 28 set2
2 3 1 27 24 29 set1
3 4 9 28 23 30 set3
2 10 2 25 26 27 set2
8 40 5 28 23 30 set3
etc...
What I need to do is select part of the data based on the
identifier. Is there an easy way of doing this that I have missed in
the egs, documentation and archives? I have tried all the
combinations/settings for categorise and lookup (after converting
the group to a field) I can think of in order to extract the data
but have yet to get it right. I think this must be a pretty normal
task so I am assuming there is something I have over looked.
Thanks for the help
Peter
The example XYZs imply this is scattered data (not a grid). Is that
correct? If not, don't go any further with this solution cause it
won't apply.
First, get out Print, set it to "rd" and use it frequently to follow
the discussion.
If so, use ImportSpreadsheet (ISS) to read in the data. Although ISS
has a built-in call to Categorize, it has been buggy in the past, so
safer is to import the data then pass it to Categorize. Tell
Categorize to operate on the column "ID" assuming you've opened up
the ISS config dialog and told it that your column names are on Line
1 (last entry in the config box as I remember and normally hidden
until you Expand and scroll down). If you don't do this, you have to
keep track of which is the auto-named "column0", etc.
Out of Categorize will come a field with (at least) 8 components: X,
Y, Z, d1, d2, d3, ID, and 'ID lookup' (which contains 3 strings
"set1", "set2", "set3"). The "ID" component now is a byte array
containing 00, 01, and 02 to indicate which value in the lookup table
corresponds to each data row (or more correctly, "position-dependent
value"). Since the lookup table is generated by alphabetizing the
input values, I would guess that your first row above will get "ID"
of 00.
ISS has a side-effect, sometimes annoying, sometimes invaluable, that
it assumes your data rows are, like beads on a string, connected by
"line" connections, one to the next (you cannot import a 3D grid with
ISS!). So if you display your data with ShowConnections, you'd get a
wacky line that traverses space like a ball of string (I'm guessing,
in this case). Use Remove to get rid of these connections before
performing the next step. To restore what you call XYZ to the
"positions" component (which ISS has declared to be a path array with
origin 1, delta 1, count nrows), use 3 instances of Mark in parallel
to mark "X", "Y", "Z", then Compute(float([a,b,c]) (the outputs of
the 3 Marks are made into a float 3-vector) and Unmark("positions").
Now use Mark("ID"), followed by Include. Set Include range to contain
only the subset of ID values (remember they are byte lookup values
now) of interest. So Include(min=0, max = 0.5) will pass set1 but not
the others. After Include, use Unmark(unspecified name) which will
clear the "marked status" of the field and permit you to immediately
Mark("data1") or whatever you like, for Autoglyph, for example.
Thus, you use ID to mask, but data1 to display at your 3D scattered
positions in space.
If that's not what you wanted, then you didn't explain the problem
clearly. (:-)
Chris Pelkie
Vice President/Scientific Visualization Producer
Conceptual Reality Presentations, Inc.
30 West Meadow Drive
Ithaca, NY 14850
[EMAIL PROTECTED]
