| Re: printing floats/integer [message #24096] |
Thu, 08 March 2001 14:44 |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
"Sean Heukels" <sean77=cuthere=@dds.nl> writes:
> I wrote a small module for integers. The variable is formatted and returned.
>
> "1 " returns as "1"
>
> Now this doesn't work with floats. for example if I want to print "1.22221"
> I
> dont want to see it as "1.222221000000000000000" or
> "1.222222 "
>
> Does anyone know how I can solve this ??
The first thing to understand is that there are many real numbers
which cannot be represented in floating point number system of
computers. That's just the facts.
Second, if you simply don't want spaces around your numbers, why not
try STRTRIM?
One of the most sophisticated answers might be found in the following
paper by Burger and Dybvig, "Printing Floating-Point Numbers Quickly
and Accurately:"
http://citeseer.nj.nec.com/28233.html
Unfortunately that's probably overkill, and too hard to implement in
IDL.
You've seen Bob S's implementation. I have a program called INPUTFORM
on my web page which prints a number as a string, such that it can be
read again by the IDL parser. Like this:
IDL> print, inputform(1.22221)
1.22221E
You see the "E" indicates that it is a single-precision floating point
number, to distinguish it unambiguously from double precision or
integers. Just as Bob's code is not too pretty (sorry Bob), neither
is the code in INPUTFORM. Basically it tries printing the number with
both the "G" and "D" output formats and takes whichever is shorter.
Craig
INPUTFORM can be found at
http://cow.physics.wisc.edu/~craigm/idl/idl.html
--
------------------------------------------------------------ --------------
Craig B. Markwardt, Ph.D. EMAIL: craigmnet@cow.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
------------------------------------------------------------ --------------
|
|
|
|
| Re: printing floats/integer [message #24108 is a reply to message #24096] |
Thu, 08 March 2001 09:27  |
R.G.S.
Messages: 46
Registered: September 2000
|
Member |
|
|
Sean Heukels <sean77=cuthere=@dds.nl> wrote in message
news:9881p0$55o$1@newshost.accu.uu.nl...
> I wrote a small module for integers. The variable is formatted and
returned.
>
> "1 " returns as "1"
>
> Now this doesn't work with floats. for example if I want to print
"1.22221"
> I
> dont want to see it as "1.222221000000000000000" or
> "1.222222 "
>
> Does anyone know how I can solve this ??
>
> Greets Sean
Please see the following code.
Note that there is a main level example code at the bottom of the function.
Using a = 123121.22222200000d
I get the following output:
Normal print command:
123121.22
Print command with GetFormat:
123121.222222
Cheers,
bob stockwell
stockwell@co-ra.com
function getformatstring,number
; this should be big enough to show everything
showallformatcommmand = '(f50.25)'
nsize = size(number,/st)
case nsize.type of
;0: Undefined
;1: Byte
;2: Integer
;3: Longword integer
4: begin; Floating point
s = STRING(FORMAT=showallformatcommmand, number)
s = strtrim(s,1) ; remove leading blanks
slen = strlen(s)
dpos=strpos(s,'.') ; find decimal posiiton
if dpos eq -1 then begin
; this shouldn't happen
print,'Error: no decimal place found'
formatstring = ''
endif else begin
; chop off lagging 0s
zpos=0
doflag=1
while doflag do begin ; find how many lagging zeros there are
char = strmid(s,zpos,1,/reverse)
if char eq '0' then begin
zpos=zpos+1
endif else begin
doflag=0
endelse
endwhile
endelse
end
5: begin; Double-precision floating
s = STRING(FORMAT=showallformatcommmand, number)
s = strtrim(s,1) ; remove leading blanks
slen = strlen(s)
print,'slen:',slen
print,s
dpos=strpos(s,'.') ; find decimal posiiton
if dpos eq -1 then begin
; this shouldn't happen
print,'Error: no decimal place found'
formatstring = ''
endif else begin
; chop off lagging 0s
zpos=0
doflag=1
while doflag do begin ; find how many lagging zeros there are
char = strmid(s,zpos,1,/reverse)
if char eq '0' then begin
zpos=zpos+1
endif else begin
doflag=0
endelse
endwhile
endelse
end
;6:; Complex floating
;7: String
;8: Structure
;9:; Double-precision complex
;10: Pointer
;11: Object reference
;12:; Unsigned Integer
;13:; Unsigned Longword Integer
;14:; 64-bit Integer
;15:; Unsigned 64-bit Integer
else: begin
print,'Error: Number not a supported type.'
return,'' ; null string
endelse
endcase
d =slen-dpos-zpos-1
w = dpos+d+1 ; 1 for the decimal place
formatstring = strcompress("(f"+string(w)+'.'+string(d)+")",/rem)
return,formatstring
end
; ****** main level code here
a = 123121.22222200000d
;must become 1.22222 without leading or trailing spaces.
b= 1.00
print,'Normal print command:'
print,a
print,'Print command with GetFormat:'
print,a,format=getformatstring(a)
end
|
|
|
|
| Re: printing floats/integer [message #24110 is a reply to message #24108] |
Thu, 08 March 2001 09:01 |
Christopher W. O'Dell
Messages: 20
Registered: February 2001
|
Junior Member |
|
|
This problem has been around about as long as computers I think. I used
to have an old pascal routine that would tell you what format code you
needed for a given number, assuming you wanted to see the whole number,
but not anymore more than is necessary (for instance, 8.6). Has anyone
written such a function for IDL?
You could probably piece together what you need using this function I just
found on the IDL Libraries Browser at Washington (see below).
Good luck,
Chris
;----------------------------------------------------------- -------
function sigfig, range
;
;+
; NAME:
; SIGFIG
;
; PURPOSE:
; This function will return the number of significant figures in
; the value "range"
;
; CATEGORY:
; utilities
;
; CALLING SEQUENCE:
; result = sigfig(range)
; INPUTS:
; range = range for the significance, may be an array.
;
; OUTPUTS:
; result = the number of significant figures expressed in base 10,
; Example: sigfig(1000) = 3,
; sigfig(0.01) = -2
;
; COMMON BLOCKS:
; none.
; SIDE EFFECTS:
; none.
; MODIFICATION HISTORY:
; Written by: Trevor Harris, Physics Dept., University of Adelaide,
; July, 1990.
|
|
|
|
| Re: printing floats/integer [message #24111 is a reply to message #24110] |
Thu, 08 March 2001 08:00 |
Klaus Scipal
Messages: 45
Registered: November 1997
|
Member |
|
|
the only solution to your problem I would come up with is doing it in a loop
recursively but i am afraid this is not very elegant
klaus
Sean Heukels <sean77=cuthere=@dds.nl> wrote in message
news:9884hb$coo$1@newshost.accu.uu.nl...
>
> What if I dont know that the float will
> be 8 in length and has 6 decimals ??
>
> How can we put this in a subroutine, that handles
> a variable (e.g. PRO FORMAT, var1) and returns
> the right way to print ??
>
> For the int I use
>
> PRO FORMAT, var
> which=SIZE(var, /TYPE)
> if (which eq 2) then begin
> le=FLOOR(ALOG10(var)) + 1
> RETURN, STRING(var, FORMAT='(I'+STRTRIM(STRING(le))+')'
> endif else if (which eq 4) then begin
>
> How do I do this for a float >>>??
> 1.22222200000 must become 1.22222 without leading or trailing spaces.
> And in the best scenario 1.00 doesn;t become 1., but 1.0 ....
>
>
>
> Sean
>
>
> Klaus Scipal <kscipal@ipf.tuwien.ac.at> schreef in berichtnieuws
> 9882vb$sbi$1@news.tuwien.ac.at...
>> you can use the the string command in connection with format
>>
>> for example
>> IDL> x=1.222221000000000000000
>> IDL> print, string(format='(f8.6)',x)
>> 1.222221
>>
>> klaus
>>
>> Sean Heukels <sean77=cuthere=@dds.nl> wrote in message
>> news:9881p0$55o$1@newshost.accu.uu.nl...
>>> I wrote a small module for integers. The variable is formatted and
>> returned.
>>>
>>> "1 " returns as "1"
>>>
>>> Now this doesn't work with floats. for example if I want to print
>> "1.22221"
>>> I
>>> dont want to see it as "1.222221000000000000000" or
>>> "1.222222 "
>>>
>>> Does anyone know how I can solve this ??
>>>
>>> Greets Sean
>>>
>>>
>>
>>
>
>
|
|
|
|
| Re: printing floats/integer [message #24113 is a reply to message #24111] |
Thu, 08 March 2001 07:16 |
Paul van Delst
Messages: 364
Registered: March 1997
|
Senior Member |
|
|
Sean Heukels wrote:
>
> What if I dont know that the float will
> be 8 in length and has 6 decimals ??
You, as the user of the number, have to specify something about what you want for the
output format. E.g. maximum length, how many decimal places, etc. IDL, and computers in
general, have absolutely no intelligence - that's where the user comes in. :o) If you
don't know what the size of the number will be, use exponential format.
> How can we put this in a subroutine, that handles
> a variable (e.g. PRO FORMAT, var1) and returns
> the right way to print ??
You haven't defined what "right" is in a way a program can understand it.
--
Paul van Delst A little learning is a dangerous thing;
CIMSS @ NOAA/NCEP Drink deep, or taste not the Pierian spring;
Ph: (301)763-8000 x7274 There shallow draughts intoxicate the brain,
Fax:(301)763-8545 And drinking largely sobers us again.
paul.vandelst@noaa.gov Alexander Pope.
|
|
|
|
| Re: printing floats/integer [message #24116 is a reply to message #24113] |
Thu, 08 March 2001 06:21 |
Sean Heukels
Messages: 25
Registered: November 1999
|
Junior Member |
|
|
What if I dont know that the float will
be 8 in length and has 6 decimals ??
How can we put this in a subroutine, that handles
a variable (e.g. PRO FORMAT, var1) and returns
the right way to print ??
For the int I use
PRO FORMAT, var
which=SIZE(var, /TYPE)
if (which eq 2) then begin
le=FLOOR(ALOG10(var)) + 1
RETURN, STRING(var, FORMAT='(I'+STRTRIM(STRING(le))+')'
endif else if (which eq 4) then begin
How do I do this for a float >>>??
1.22222200000 must become 1.22222 without leading or trailing spaces.
And in the best scenario 1.00 doesn;t become 1., but 1.0 ....
Sean
Klaus Scipal <kscipal@ipf.tuwien.ac.at> schreef in berichtnieuws
9882vb$sbi$1@news.tuwien.ac.at...
> you can use the the string command in connection with format
>
> for example
> IDL> x=1.222221000000000000000
> IDL> print, string(format='(f8.6)',x)
> 1.222221
>
> klaus
>
> Sean Heukels <sean77=cuthere=@dds.nl> wrote in message
> news:9881p0$55o$1@newshost.accu.uu.nl...
>> I wrote a small module for integers. The variable is formatted and
> returned.
>>
>> "1 " returns as "1"
>>
>> Now this doesn't work with floats. for example if I want to print
> "1.22221"
>> I
>> dont want to see it as "1.222221000000000000000" or
>> "1.222222 "
>>
>> Does anyone know how I can solve this ??
>>
>> Greets Sean
>>
>>
>
>
|
|
|
|
| Re: printing floats/integer [message #24117 is a reply to message #24116] |
Thu, 08 March 2001 05:57 |
Klaus Scipal
Messages: 45
Registered: November 1997
|
Member |
|
|
you can use the the string command in connection with format
for example
IDL> x=1.222221000000000000000
IDL> print, string(format='(f8.6)',x)
1.222221
klaus
Sean Heukels <sean77=cuthere=@dds.nl> wrote in message
news:9881p0$55o$1@newshost.accu.uu.nl...
> I wrote a small module for integers. The variable is formatted and
returned.
>
> "1 " returns as "1"
>
> Now this doesn't work with floats. for example if I want to print
"1.22221"
> I
> dont want to see it as "1.222221000000000000000" or
> "1.222222 "
>
> Does anyone know how I can solve this ??
>
> Greets Sean
>
>
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
