| Re: Array elements as named variables? [message #2360] |
Tue, 28 June 1994 10:37  |
jimbrakefd
Messages: 4
Registered: March 1994
|
Junior Member |
|
|
In article <2un02p$lm1@lace.Colorado.EDU>, msegur@newton (The
Ethereal Knight) writes:
> Is it possible to used an element of an array as a named
variable
> in order for a procedure to store a value in it? When I try to do
this, I
> get an error telling me that the expression must be a named
variable, but
> it seems like a reasonable thing to want to do. I know that I can
use a
> dummy variable to store the value, but this seems awfully clumsy. Am
I
> missing something?
Of course one can pass the subscripted array as a text string for
EXECUTE
(.i.e. EXECUTE("A[2,5]" + "=" + "result")
Jim Brakefield
|
|
|
|
| Re: Array elements as named variables? [message #2361 is a reply to message #2360] |
Tue, 28 June 1994 11:16  |
schrick
Messages: 2
Registered: June 1994
|
Junior Member |
|
|
The Ethereal Knight (msegur@newton) wrote:
: Is it possible to used an element of an array as a named variable
: in order for a procedure to store a value in it?
No, but it is possible to pass whole arrays between procedures:
pro ncdf_diminq, cdfid, i, dname, dsize
dname(i) = "testing 123"
dsize(i) = 100
end
;main
d_name = make_array(glob.ndims, /STRING)
d_size = make_array(glob.ndims, /LONG)
ncdf_diminq, cdfid, i, d_name, d_size
end
Mike
|
|
|
|
| Re: Array elements as named variables? [message #2368 is a reply to message #2360] |
Mon, 27 June 1994 15:11 |
landers
Messages: 45
Registered: May 1993
|
Member |
|
|
In article <2un02p$lm1@lace.Colorado.EDU>, msegur@newton (The Ethereal Knight) writes:
|> Is it possible to used an element of an array as a named variable
|> in order for a procedure to store a value in it?
[snip]
Nope, can't do it. Seems reasonable that you should be able to, since it
points to a single memory location (at least at the time of the CALL, if not
the RETURN, since the varaible referenced could change during the subroutine
processing!...), and you could have used it on the left-hand-side of an
assignment. But...
When you call a procedure, IDL/WAVE decides whether to pass a parameter by
value or by reference (memory address). There is no consideration given to
making assignments upon a RETURN. Only things passed by reference can be
changed, since the location of the actual variable in memory is the same in
the caller and the subroutine.
The reasoning used to decide how to pass a parameter is if it's an expression
or not. Check this by doing HELP (IDL) or INFO (WAVE) like this:
WAVE> info,a
A FLOAT = Array(10)
WAVE> info,a(0)
<Expression> FLOAT = 0.00000
^^^^^^^^^^^^
You want it to decide based on if you could use a reference on the left-hand
side of an expression. Not me. The way it is, I can pass an array to a
procedure like this:
Some_Procedure, a
or like this:
Some_Procedure, a(*)
The procedure gets the same data, but in the second case, I can insure that my
data won't be changed behind my back. Just a matter of perspective.
;Dave
|
|
|
|
| Re: Array elements as named variables? [message #2370 is a reply to message #2368] |
Mon, 27 June 1994 10:16 |
thompson
Messages: 584
Registered: August 1991
|
Senior Member |
|
|
msegur@newton (The Ethereal Knight) writes:
> Is it possible to used an element of an array as a named variable
> in order for a procedure to store a value in it?
The simple answer is no. Unfortunately, the inner mechanics of IDL doesn't
allow you to do this. You have to do something like this
d_name = make_array(glob.ndims, /STRING)
d_size = make_array(glob.ndims, /LONG)
ncdf_diminq, cdfid, i, temp_name, temp_size
d_name(i) = temp_name
d_size(i) = temp_size
(assuming that the last two parameters of ncdf_diminq are output parameters.
Bill Thompson
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
