| The limits of !NULL [message #83938] |
Wed, 17 April 2013 19:52 |
Jeremy Bailin
Messages: 618
Registered: April 2008
|
Senior Member |
|
|
Having only recently started switching to IDL8 thinking, I decided today
to try out the WHERE(/NULL) functionality. And the first thing I tried
to do failed miserably.
I think the fundamental problem is that !NULL doesn't seem to be at all
useful on the RHS of an equation. What I wanted was to shorten:
a = where(array eq value, na)
if na gt 0 then begin
data1[a] = -data1[a]
data2[a] = -data2[a]
endif
to:
a = where(array eq value, /null)
data1[a] = -data1[a]
data2[a] = -data2[a]
since I thought the general idea of this functionality was so that you
no longer had to test for no matches. However, when the condition is not
met, it fails miserably:
IDL> data1[a] = -data1[a]
% Variable is undefined: <UNDEFINED>.
IDL> help, a
A UNDEFINED = !NULL
If a is used only on the LHS, it works fine:
IDL> data1[a] = 5
IDL>
But even if the RHS is implicit, it fails:
IDL> data1[a]++
% Variable is undefined: A.
So I guess I'm sticking with the original coding paradigm.
-Jeremy.
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
