| Re: transforming an array where some values can't [message #10896] |
Mon, 09 February 1998 00:00 |
Evilio del Rio
Messages: 17
Registered: December 1997
|
Junior Member |
|
|
On 6 Feb 1998, Joel D. Offenberg wrote:
> bowler@eisner.decus.org writes:
>
>> I'm trying to write a function that will transform an arbitrary array by taking
>> hte natural log of every element in the array.
>
>> What I have is
>
>> function xform, input
>> temp = alog(input) ; I realize I could combine these 2 statements
>> return, temp ; but it makes debugging easier
>> end
>
>> Unfortunately, I can't guarentee that there won't be some elements that are
>> zero and thus invalid arguments to alog. what's the "most efficient" way to
>> take the alog of any element that's greater than 0 and set the value of ony
>> that are less than or equal to 0 to some small value (1e-7 for example)?
>
IDL softly handles any math exception (IEEE standard) so you don't need to
worry about the small/negative values for alog() :
IDL> print,alog(0.0)
-inf
IDL> print,alog(-1.0)
nan0x2000000
You just need to handle special case with the function FINITE(). However,
many of the standard plot/calculation routines can treat naN values as
missing data, try for example:
IDL> plot,alog(randomn(seed,100))
Cheers,
____________________________________________________________ ________
Evilio Jose del Rio Silvan Institut de Ciencies del Mar
E-mail: edelrio@icm.csic.es URL: http://www.ieec.fcr.es/~evilio/
"Anywhere you choose,/ Anyway, you're gonna lose"- Mike Oldfield
|
|
|
|
| Re: transforming an array where some values can't [message #10897 is a reply to message #10896] |
Sun, 08 February 1998 00:00  |
alpha
Messages: 49
Registered: September 1996
|
Member |
|
|
offenbrg@fondue.gsfc.nasa.gov (Joel D. Offenberg) writes:
> bowler@eisner.decus.org writes:
>> What I have is
>> function xform, input
>> temp = alog(input) ; I realize I could combine these 2 statements
>> return, temp ; but it makes debugging easier
>> end
>> Unfortunately, I can't guarentee that there won't be some elements that are
>> zero and thus invalid arguments to alog. what's the "most efficient" way to
>> take the alog of any element that's greater than 0 and set the value of ony
>> that are less than or equal to 0 to some small value (1e-7 for example)?
> I think what you want is this:
> function xform, input
> temp = alog(input > 1e-7)
> return, temp
> end
> The ">" operator in this case sets a floor on the values of "input"...any
> value less than 1e-7 is replaced with 1e-7.
RSI told us, that the alo10 error was corrected in release 5.1!
so wait
Panther
--
Panther in the Jungle __..--''``\--....___ _..,_
-BELIEVE AND DECEIVE- _.-' .-/"; ` ``<._ ``-+'~=.
http://www.ang-physik _.-' _..--.'_ \ `(^) )
.uni-kiel.de/~hendrik ((..-' (< _ ;_..__ ; `'
|
|
|
|
| Re: transforming an array where some values can't [message #10901 is a reply to message #10897] |
Fri, 06 February 1998 00:00 |
offenbrg
Messages: 31
Registered: August 1993
|
Member |
|
|
bowler@eisner.decus.org writes:
> I'm trying to write a function that will transform an arbitrary array by taking
> hte natural log of every element in the array.
> What I have is
> function xform, input
> temp = alog(input) ; I realize I could combine these 2 statements
> return, temp ; but it makes debugging easier
> end
> Unfortunately, I can't guarentee that there won't be some elements that are
> zero and thus invalid arguments to alog. what's the "most efficient" way to
> take the alog of any element that's greater than 0 and set the value of ony
> that are less than or equal to 0 to some small value (1e-7 for example)?
I think what you want is this:
function xform, input
temp = alog(input > 1e-7)
return, temp
end
The ">" operator in this case sets a floor on the values of "input"...any
value less than 1e-7 is replaced with 1e-7.
Good luck,
Joel
--
"...And I am unanimous in this" - Mrs. Slocumbe
------------------------------------------------------------ ---------------
| Joel D Offenberg | Joel.D.Offenbrg.1@gsfc.nasa.gov |
| Hughes STX, NASA/GSFC/LASP & STScI | UIT, NGST programmer & sysadmin |
|
|
|
|
| Re: transforming an array where some values can't [message #10907 is a reply to message #10901] |
Thu, 05 February 1998 00:00 |
S Penzes
Messages: 4
Registered: February 1998
|
Junior Member |
|
|
How about
function xform, input
temp = alog(input)
temp(where(finite(temp) EQ 0))=1e-7
return, temp
end
Works for me.
bowler@eisner.decus.org wrote:
> I'm trying to write a function that will transform an arbitrary array by taking
> hte natural log of every element in the array.
>
> What I have is
>
> function xform, input
> temp = alog(input) ; I realize I could combine these 2 statements
> return, temp ; but it makes debugging easier
> end
>
> Unfortunately, I can't guarentee that there won't be some elements that are
> zero and thus invalid arguments to alog. what's the "most efficient" way to
> take the alog of any element that's greater than 0 and set the value of ony
> that are less than or equal to 0 to some small value (1e-7 for example)?
>
> TIA,
> Bruce
--
Steven Penzes (Steven.nospamPenzes@dres.dnd.ca)
Note: remove "nospam" from Reply-To and .signature
|
|
|
|
| Re: transforming an array where some values can't [message #10908 is a reply to message #10907] |
Thu, 05 February 1998 00:00 |
meron
Messages: 51
Registered: July 1995
|
Member |
|
|
In article <1998Feb5.094807.1@eisner>, bowler@eisner.decus.org writes:
> I'm trying to write a function that will transform an arbitrary array by taking
> hte natural log of every element in the array.
>
> What I have is
>
> function xform, input
> temp = alog(input) ; I realize I could combine these 2 statements
> return, temp ; but it makes debugging easier
> end
>
> Unfortunately, I can't guarentee that there won't be some elements that are
> zero and thus invalid arguments to alog. what's the "most efficient" way to
> take the alog of any element that's greater than 0 and set the value of ony
> that are less than or equal to 0 to some small value (1e-7 for example)?
>
Just do
function xform, input
eps = ..some small value here, can make it input parameter too ..
temp = alog(input > eps) ; I realize I could combine these 2 statements
return, temp ; but it makes debugging easier
end
Mati Meron | "When you argue with a fool,
meron@cars.uchicago.edu | chances are he is doing just the same"
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
