| Re: Empty arrays? [message #63437] |
Mon, 10 November 2008 14:56 |
Michael Galloy
Messages: 1114
Registered: April 2006
|
Senior Member |
|
|
On Nov 10, 3:35 pm, Spon <christoph.b...@gmail.com> wrote:
> For these sorts of cases, I tend to use empty pointers. Here's an
> example using pointers to provide the functionality I think you're
> looking for:
>
> var = ptr_new(/allocate_heap)
> help, *var
> IDL> <PtrHeapVar1> UNDEFINED = <Undefined>
>
> This article by JD Smith on David Fanning's website should help get
> you started on pointers:http://www.dfanning.com/misc_tips/pointers.html
Yes, empty pointers work well for these type of things, but there are
still two issues with them in this situation:
1. you still have to check to see if *var is undefined to determine
if you are appending to or creating an array.
2. repeatedly appending to an array is inefficient
A better solution is to create a reasonable sized array to begin with.
Then fill in values and keep track of how many are filled in. If the
array fills up, create a new bigger one and copy the values over. Of
course, this is even more of a hassle with bookkeeping, so I made an
object that does this for me, see:
http://michaelgalloy.com/2006/04/24/collection-package-mgarr aylist.html
Of course, it would be nicer if IDL just allowed empty arrays, but
there would be a lot of backward compatibilities if that were the
case.
Mike
--
www.michaelgalloy.com
Tech-X Corporation
Associate Research Scientist
|
|
|
|
| Re: Empty arrays? [message #63438 is a reply to message #63437] |
Mon, 10 November 2008 14:35  |
Spon
Messages: 178
Registered: September 2007
|
Senior Member |
|
|
On Nov 10, 10:07 pm, Demitri wrote:
> Hello,
>
> I'm starting to dig into IDL programming, and one thing I've found
> particularly annoying is the lack of support for empty arrays. Although
> I'm using IDL's OO capabilities, I find that I'm spending a lot of time
> testing for return results where an empty array would be much more
> suitable.
>
> If I'm building an array that will have zero or more entries, I create
> a variable, "insert" a dummy value, go through the code adding entries
> based on whatever logic, then pop the top dummy value. But even then I
> have to check that more than one entry is there so I dont throw an
> error with a
>
> return, a[1:*]
>
> This is very basic functionality. Does anyone have a particularly
> elegant... well, I'd say solution, but anything is obviously a
> workaround. I don't want to fight the language, but I get the feeling
> that IDL's philosophy is "yeah, you do it yourself. I can't be
> bothered."
>
> Cheers,
>
> Demitri
Hi Demitri,
For these sorts of cases, I tend to use empty pointers. Here's an
example using pointers to provide the functionality I think you're
looking for:
var = ptr_new(/allocate_heap)
help, *var
IDL> <PtrHeapVar1> UNDEFINED = <Undefined>
This article by JD Smith on David Fanning's website should help get
you started on pointers:
http://www.dfanning.com/misc_tips/pointers.html
Good luck!
Chris
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
