| Re: V6.0 indexing change [message #36176] |
Wed, 13 August 2003 23:20 |
R.Bauer
Messages: 1424
Registered: November 1998
|
Senior Member |
|
|
Wayne Landsman wrote:
> I haven't installed V6.0 yet myself but I have been informed by Mike
> Corcoran of a change that can break some existing code.
>
> Formerly, if one had a structure, say a = {x:indgen(10)} which was
> accessed by a 1 element vector, say i = [3], then the result a.x[i]
> would be a scalar. In V6.0, however, the result is now a 1 element
> vector.
>
> I have to say that the V6.0 method makes more more sense -- subscripting
> with a 1 element vector should return a 1 element vector. But it can
> break existing code because logical tests such as
>
> if a.x[i] EQ 3 then stop
>
> will no longer work. The fix is simple enough ( set i = i[0] first )
> but users might want to fix the code before upgrading to V6.0.
Could this change be interpreted that's trailing dimensions 1 sometimes
later not more removed automaticly.
a=make_Array(4,1)
IDL> help,a
A FLOAT = Array[4]
Reimar
>
> --Wayne Landsman landsman@mpb.gsfc.nasa.gov
--
Forschungszentrum Juelich
email: R.Bauer@fz-juelich.de
http://www.fz-juelich.de/icg/icg-i/
============================================================ ======
a IDL library at ForschungsZentrum Juelich
http://www.fz-juelich.de/icg/icg-i/idl_icglib/idl_lib_intro. html
|
|
|
|
| Re: V6.0 indexing change [message #36186 is a reply to message #36176] |
Tue, 12 August 2003 18:51  |
news.verizon.net
Messages: 47
Registered: August 2003
|
Member |
|
|
"Mark Hadfield" <m.hadfield@niwa.co.nz> wrote in message
news:bhc1o5$kl2$3@newsreader.mailgate.org...
>
> Still, Wayne's warning is correct: there may be cases where the new,
> improved structure indexing in 6.0 will break code.
>
Yes, the problem evidently was not with a logical test, but with a case
statement, e.g.
case a.x[i] of
1: print,'hi'
else: print,'bye'
endcase
where a.x[i] is required to be a scalar and not a 1 element vector. --Wayne
P.S. The V6.0 incompatibility occured in the program tbget.pro in the IDL
Astro library.
|
|
|
|
| Re: V6.0 indexing change [message #36188 is a reply to message #36186] |
Tue, 12 August 2003 17:45 |
Mark Hadfield
Messages: 783
Registered: May 1995
|
Senior Member |
|
|
Andrew Cool wrote:
> Wayne Landsman <landsman@mpb.gsfc.nasa.gov> wrote in message news:<3F39035C.8070409@mpb.gsfc.nasa.gov>...
>
>> I haven't installed V6.0 yet myself but I have been informed by Mike
>> Corcoran of a change that can break some existing code.
>>
>> Formerly, if one had a structure, say a = {x:indgen(10)} which was
>> accessed by a 1 element vector, say i = [3], then the result a.x[i]
>> would be a scalar. In V6.0, however, the result is now a 1 element
>> vector.
>>
>> I have to say that the V6.0 method makes more more sense -- subscripting
>> with a 1 element vector should return a 1 element vector. But it can
>> break existing code because logical tests such as
>>
>> if a.x[i] EQ 3 then stop
>>
>> will no longer work. The fix is simple enough ( set i = i[0] first )
>> but users might want to fix the code before upgrading to V6.0.
>>
>
> You sure about this? Yes, a 1 element vector is now returned, but
> your example
>
> if a.x[i] EQ 3 then stop
>
> works just fine for me under the 6.0 Beta (Win2000), as does
>
> if a.x[i[0]] EQ 3 then stop
And for me under 6.0 final (Win 2000). But then this...
IDL> if [1] eq 1 then print, 'comparison is OK'
...works (ie prints out the message) under IDL 6.0 and 5.6. In other
words an if statement doesn't seem to care if its condition is a scalar
or a 1-element array. News to me!
Still, Wayne's warning is correct: there may be cases where the new,
improved structure indexing in 6.0 will break code.
--
Mark Hadfield "Ka puwaha te tai nei, Hoea tatou"
m.hadfield@niwa.co.nz
National Institute for Water and Atmospheric Research (NIWA)
Gratuitous text to fool the news server. Gratuitous text to fool the
news server. Gratuitous text to fool the news server. Gratuitous text to
fool the news server. Gratuitous text to fool the news server.
Gratuitous text to fool the news server. Gratuitous text to fool the
news server. Gratuitous text to fool the news server. Gratuitous text to
fool the news server. Gratuitous text to fool the news server.
Gratuitous text to fool the news server. Gratuitous text to fool the
news server. Gratuitous text to fool the news server. Gratuitous text to
fool the news server. Gratuitous text to fool the news server.
Gratuitous text to fool the news server. Gratuitous text to fool the
news server. Gratuitous text to fool the news server. Gratuitous text to
fool the news server. Gratuitous text to fool the news server.
|
|
|
|
| Re: V6.0 indexing change [message #36189 is a reply to message #36188] |
Tue, 12 August 2003 16:04 |
andrew.cool
Messages: 47
Registered: July 2003
|
Member |
|
|
Wayne Landsman <landsman@mpb.gsfc.nasa.gov> wrote in message news:<3F39035C.8070409@mpb.gsfc.nasa.gov>...
> I haven't installed V6.0 yet myself but I have been informed by Mike
> Corcoran of a change that can break some existing code.
>
> Formerly, if one had a structure, say a = {x:indgen(10)} which was
> accessed by a 1 element vector, say i = [3], then the result a.x[i]
> would be a scalar. In V6.0, however, the result is now a 1 element
> vector.
>
> I have to say that the V6.0 method makes more more sense -- subscripting
> with a 1 element vector should return a 1 element vector. But it can
> break existing code because logical tests such as
>
> if a.x[i] EQ 3 then stop
>
> will no longer work. The fix is simple enough ( set i = i[0] first )
> but users might want to fix the code before upgrading to V6.0.
>
> --Wayne Landsman landsman@mpb.gsfc.nasa.gov
Wayne,
You sure about this? Yes, a 1 element vector is now returned, but
your example
if a.x[i] EQ 3 then stop
works just fine for me under the 6.0 Beta (Win2000), as does
if a.x[i[0]] EQ 3 then stop
Andrew
DSTO, Adelaide, South Australia
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
