| structure array changed in 5.5 [message #28467] |
Wed, 12 December 2001 14:37 |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
Here's another interesting tidbit about subscripting, this one only
applying to IDL 5.5.
I always complained that extracting array values from structures can
be dangerous. Consider the following structure:
s = {x: [1]}
The tag X is a one-element vector. Under IDL 5.4 and earlier,
retrieving the value of S.X returned a *scalar*.
IDL> help, s.x
<Expression> INT = 1
This is a problem if you don't know ahead of time whether X should be
an array or scalar, because there is no way to find out.
In IDL 5.5, this behavior has changed, for the better I think.
IDL> help, s.x
<Expression> INT = Array[1]
This also works for higher dimensional arrays, as they should work.
The only problem is if you depend on the former behavior. Aaron Barth
(et al.) found a problem with my published CMPS_FORM. It relied on
the following code to work
configs = predefined(*).config
However, the CONFIG tag is a structure itself, and since structures
are always arrays in IDL, the returned value is different under
pre-5.5 and post-5.5.
pre 5.5: CONFIGS STRUCT = -> CMPS_FORM_INFO Array[6]
post 5.5: CONFIGS STRUCT = -> CMPS_FORM_INFO Array[1, 6]
You see one is a 1x6 array and one is a 6 element array. This fouled
things up down the line (now corrected).
The short answer is that if you depend on the behavior of arrays of
structures with arrays (!), then you should double check your code to
avoid the above snag.
Yours,
Craig
--
------------------------------------------------------------ --------------
Craig B. Markwardt, Ph.D. EMAIL: craigmnet@cow.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
------------------------------------------------------------ --------------
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
