|
|
| Re: Creating Variables in Programs [message #12301 is a reply to message #12195] |
Tue, 14 July 1998 00:00  |
davidf
Messages: 2866
Registered: September 1996
|
Senior Member |
|
|
Riemar Bauer (r.bauer@fz-juelich.de) writes:
> An other way to get an undefined
> variabel is b=n_elements(a)
> Yes b is well defined by 0 but a is defined as undefined.
>
> help
> % At $MAIN$
> B LONG = 0
> A UNDEFINED = <Undefined>
Uh, this *only* happens if A is undefined to start with
and is perfectly normal behavior. In fact, it is the only
way to know if keywords are undefined in an IDL procedure or
function:
IF N_Elements(keyword) EQ 0 THEN keyword = 5
This construction NEVER makes A undefined (or there is
something seriously wrong with your version of IDL):
A = 5
B = N_Elements(A)
HELP, A, B
A INT = 5
B LONG = 1
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting
E-Mail: davidf@dfanning.com
Phone: 970-221-0438
Coyote's Guide to IDL Programming: http://www.dfanning.com/
|
|
|
|
| Re: Creating Variables in Programs [message #12303 is a reply to message #12195] |
Tue, 14 July 1998 00:00 |
R. Bauer
Messages: 137
Registered: November 1996
|
Senior Member |
|
|
Martin Schultz wrote:
> Craig Markwardt wrote:
>>
>>> IDL> r=execute('a=fltarr(200)')
>>> IDL> help,a
>>> A FLOAT = Array[200]
>>>
>>
>> There is a "gotcha." in the case of a compiled procedure, the
>> variable "a" must have already been defined. The following is usually
>> sufficient:
>> [...]
>
> Huh? Here is a little program:
>
> --------------------------------------
> pro testexec,name
>
> r=execute(name+'=findgen(10)')
>
> print,r
> print,b
> return
> end
> --------------------------------------
>
> Of course, you have to call it as testexec,'b' in order to have it work
> properly ;-), but it demonstrates that you don't have to have your
> variable initialized!!
>
> But I don't really see the point of the original question: why the h...
> do you want to do this? To my knowledge, creating variables only makes
> sense if you know what to do with them afterwards - and in order to do
> something with them, you must know their name beforehand. If you want to
> export your newly created variables to the main program or some other
> procedure, you would have to proceed completely different. I would
> create a structure with
> template = { name:'', pvalue:ptr_new() }
> (or an array of these structures with replicate(...) )
>
> then manipulatge the string 'name=expression' to 'tmp=expression', store
> the 'name' field in the name tag of the structure and
> pvalue=ptr_new(tmp) will save the value.
>
> This would act as a container (sounds awfully like OOP doesn't it ?),
> and you would have to do a lot of type and error checking in any routine
> that uses the information in this structure (array). Note, that IDL
> itself would not "know" anything about your variables - but, as I said,
> it doesn't make sense if it had to.
>
> ... and don't forget to clean up your heap once a while...
>
Hi Martin,
that's not totally correct.
idl knows a lot of your variables which are defined or defined as undefined
(a=n_elements(b))
print,routine_names(/variables)
for more look in the by now obsolete routine gethelp
I am using this mechanism to create a dynamical structure where are nearby
100 names with definitions (mostly descriptions for datasets like:
experiment,PI_name, param_long_name, param_units ...) are defined. And all
of them which are defined in a program will go into a structure.
In the program I have only to define param_units='K' and later on it will
be a tag name in a structure. All whats in the structure is could be
written to somewhere e.g. netCDF.
Reimar
--
R.Bauer
Institut fuer Stratosphaerische Chemie (ICG-1)
Forschungszentrum Juelich
email: R.Bauer@fz-juelich.de
|
|
|
|
| Re: Creating Variables in Programs [message #12305 is a reply to message #12195] |
Tue, 14 July 1998 00:00 |
R. Bauer
Messages: 137
Registered: November 1996
|
Senior Member |
|
|
Craig Markwardt wrote:
> Martin Schultz <mgs@io.harvard.edu> writes:
>>
>> Perry Phillips wrote:
>>>
>>> Does anyone know of a way to create a new array under program control, ie
>>> create a string and use that string to make an array. As far as I can see
>>> this is impossible in IDL?
>>>
>>> --
>>> Perry Phillips p.phillips@mail.utexas.edu
>>
>>
>> here's a quick example
>>
>>
>> IDL> r=execute('a=fltarr(200)')
>> IDL> help,a
>> A FLOAT = Array[200]
>>
>
> There is a "gotcha." in the case of a compiled procedure, the
> variable "a" must have already been defined. The following is usually
> sufficient:
>
> A = 0
> ...
> R = EXECUTE('A=FLTARR(200)')
>
> The IDL internal compiler needs to know that "A" exists before it can
> be assigned to in an EXECUTE statement. The same applies for
> restoring variables: all the variables in the SAVE file must be
> predefined in the procedure. Assigning zero to them is fine.
>
>
I am not sure if this has changed with idl 5.1.An other way to get an undefined
variabel is b=n_elements(a)
Yes b is well defined by 0 but a is defined as undefined.
help
% At $MAIN$
B LONG = 0
A UNDEFINED = <Undefined>
Reimar
--
R.Bauer
Institut fuer Stratosphaerische Chemie (ICG-1)
Forschungszentrum Juelich
email: R.Bauer@fz-juelich.de
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
