| Is there a standard 'null' array? [message #34611] |
Fri, 28 March 2003 05:47 |
Matt Feinstein
Messages: 33
Registered: July 2002
|
Member |
|
|
This is an embarassingly elementary question, and I'm -sure- that it
has a simple, elegant, and obvious answer, but...
I often find myself wanting to do things kinda like
Data_I_want_vec = Init
...
for ix = 0 , (size(data_array))[1]-1 do $
if condition(data_array[ix]) then $
Data_I_want_vec = [Data_I_want_vec, data_array[ix]]
I realize that there is analogous code for a vectorized version of my
question, but I think the for-loop version is clearer.
The questions that arise are things like: What to use for 'Init' ?
What do I do to test for a null result? What if there's no data?
Some notes:
a) Matlab has an empty vector = [] that serves these purposes. It's
possible that my desire to program in this fashion is simply a
Matlab-ism, and I need instruction in the IDL way.
b) I do not, really and truly, need instruction in the use of the
WHERE function or in the use of vectorization. Believe me, I know all
about that. The analog of my question in the vectorized case is that
it looks like I have to surround the vectorized expression with boring
and error-prone tests for null inputs and outputs.
Is there a 'standard' way of doing this sort of thing?
Matt Feinstein
--
The Law of Polarity: The probability of wiring a battery with
the correct polarity is (1/2)^N, where N is the number of times
you try to connect it.
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
