| Re: is this a bug in IDL? [message #37376] |
Tue, 16 December 2003 02:58  |
![Wonko[3] is currently offline](theme/default/images/xoffline.png.pagespeed.ic.XRkd1fkXye.png "Wonko[3] is currently offline")
Wonko[3]
Messages: 9
Registered: February 2003
|
Junior Member |
|
|
aramisgm@hotmail.com (Aramis Martinez) wrote:
> k=0
> for i=0,8759 do begin
> x=vbs(2,i)
> if x gt 0.0 then k=k+1 & x=0 & vbs(2,i)=x
> endfor
> print, k
Which is the same as:
k=0
for i=0,8759 do begin
x=vbs(2,i)
if x gt 0.0 then k=k+1
x=0
vbs(2,i)=x
endfor
print, k
> 1) How would you vectorize this? I tried a few things that I could
> remember from my work this summer, but four months of quantum
> mechanics wipes a lot from memory :)
This should do:
index = where( vbs[ 2, lindgen( 8760 ) ] gt 0, k )
if k gt 0 then vbs[ 2, index ] = 0
[...]
> Is this a bug? as far as I can tell, the statements are logically
> identical since an &'s behavior should never vary, so what in the guts
> of IDL would make these statements be different? Is there some rule
> kicking in here that's not so obvious but makes this behavior to be
> expected? Would parenthesis around the THEN block make these identical
> in IDL? The environment is either IDL 6 on Red Hat 9 or IDL 5.1 on
> Red Hat 8.1.
Using & ist the same as wrting the next statement in another line. You
need to use begin and endif statements.
Alex
--
Alex Schuster Wonko@wonkology.org PGP Key available
alex@pet.mpin-koeln.mpg.de
|
|
|
|
| Re: is this a bug in IDL? [message #37377 is a reply to message #37376] |
Tue, 16 December 2003 03:09  |
Ben Panter
Messages: 102
Registered: July 2003
|
Senior Member |
|
|
Aramis Martinez wrote:
> k=0
> for i=0,8759 do begin
> x=vbs(2,i)
> if x gt 0.0 then k=k+1 & x=0 & vbs(2,i)=x
> endfor
> print, k
>
> k=0
> for i=0,8759 do begin
> x=vbs(2,i)
> ;; version A
> if x gt 0.0 then x=0 & k=k+1 & vbs(2,i)=x
> ;; version B
> if x gt 0.0 then k=k+1 & x=0 & vbs(2,i)=x
> endfor
> print, k
>
I'm no expert, but I think that possibly your code is using & in the
wrong way - as far as I understand it (correct me if I'm wrong guys!!) &
means "complete the previous command then do the thing after me" - so
your code actually reads:
k=0
for i=0,8759 do begin
x=vbs(2,i)
if x gt 0.0 then x=0
k=k+1
vbs(2,i)=x
endfor
or
k=0
for i=0,8759 do begin
x=vbs(2,i)
if x gt 0.0 then k=k+1
x=0
vbs(2,i)=x
endfor
To get the behaviour that you want I think you need another begin...end set:
k=0
for i=0,8759 do begin
x=vbs(2,i)
if x gt 0.0 then begin
k=k+1
x=0
vbs(2,i)=x
end
endfor
--
Ben Panter, Edinburgh
My name (no spaces)@bigfoot which is a com.
|
|
|
|
| Re: is this a bug in IDL? [message #37463 is a reply to message #37376] |
Tue, 16 December 2003 13:17 |
aramisgm
Messages: 4
Registered: December 2003
|
Junior Member |
|
|
I've used index, but I'd forgotten about it. Thanks for the reminder.
Ah, the & means for each iteration do an if-then block, then do the
two following statements (as opposed to tying together the three
statements as the then part of the if-then block) -- the scope for the
& is slightly different than in, say, Java. However, now there is a
new problem. The entire array should get zeroed out if the snippet
below is what is happening. . . (five minutes and a phone call later)
indeed it is! To me this seems to be a non-standard usage of &, but
that's what I get for being trained in c and its descendants :+) The
three statements go into a do-begin-end block, problem solved.
Thank you gentlemen. Enlightenment received.
Aramis Martinez
Wonko@wonkology.org (Alex Schuster) wrote in message news:<8zysEJaud8B@wonkology.org>...
> aramisgm@hotmail.com (Aramis Martinez) wrote:
>
>> k=0
>> for i=0,8759 do begin
>> x=vbs(2,i)
>> if x gt 0.0 then k=k+1 & x=0 & vbs(2,i)=x
>> endfor
>> print, k
>
> Which is the same as:
>
> k=0
> for i=0,8759 do begin
> x=vbs(2,i)
> if x gt 0.0 then k=k+1
> x=0
> vbs(2,i)=x
> endfor
> print, k
>
>
>> 1) How would you vectorize this? I tried a few things that I could
>> remember from my work this summer, but four months of quantum
>> mechanics wipes a lot from memory :)
>
> This should do:
>
> index = where( vbs[ 2, lindgen( 8760 ) ] gt 0, k )
> if k gt 0 then vbs[ 2, index ] = 0
>
>
> [...]
>
>> Is this a bug? as far as I can tell, the statements are logically
>> identical since an &'s behavior should never vary, so what in the guts
>> of IDL would make these statements be different? Is there some rule
>> kicking in here that's not so obvious but makes this behavior to be
>> expected? Would parenthesis around the THEN block make these identical
>> in IDL? The environment is either IDL 6 on Red Hat 9 or IDL 5.1 on
>> Red Hat 8.1.
>
> Using & ist the same as wrting the next statement in another line. You
> need to use begin and endif statements.
>
> Alex
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
