| Re: significant figures function? [message #23882] |
Mon, 26 February 2001 18:33 |
Med Bennett
Messages: 109
Registered: April 1997
|
Senior Member |
|
|
Thanks Ben and Martin! I had already kluged together another solution using a
routine I found using the search engine I found using
http://www.astro.washington.edu/deutsch/idl/htmlhelp/index.h tml (very handy
tool!) by searching for 'significant' called strnsignif.pro
( http://www.astro.washington.edu/deutsch-bin/getpro/library03 .html?STRNSIGNIF).
Wrapping this with a loop and converting back to float did the trick. Less
than perfect, but in the consulting biz, whatever can be found quickly beats
an elegant solution!
Med
Martin Downing wrote:
> "Med Bennett" <mbennett@indra.com> wrote in message
> news:3A9ACD30.E0F3F5D9@indra.com...
>> Has anyone written a function that returns the input value or array with
>> a specified number of significant digits? I have lloked at the various
>> IDL libraries on the web but did not come up with what I'm after.
>> Thanks in advance for any pointers -
>>
>
> this should do it
>
> function fix_digits, num, digits
> expon = 1.0
> while num/expon gt 1 do expon = expon * 10
> fix_val = num/expon
> fstring = string(digits+2, digits, format = '("(f",i0.0,".",i0.0,")")')
> reads, string(fix_val, format = fstring), fix_val
> return, fix_val * expon
> end
>
>> print, fix_digits(!pi, 3)
> 3.14000
>
> Martin
>
> ----------------------------------------
> Mr. Martin Downing,
> Clinical Research Physicist,
> Orthopaedic RSA Research Centre,
> Woodend Hospital,
> Aberdeen, AB15 6LS.
> m.downing@abdn.ac.uk
>
> Martin
|
|
|
|
| Re: significant figures function? [message #23884 is a reply to message #23882] |
Mon, 26 February 2001 17:36  |
Martin Downing
Messages: 136
Registered: September 1998
|
Senior Member |
|
|
"Ben Tupper" <pemaquidriver@tidewater.net> wrote in message
news:3A9AEA4B.E6169330@tidewater.net...
> You try the following.
>
> ;------start
> ;+
> ; NAME:
> ; ROUNDPV
> ;
hm, I got funny results with this (as i did with my other attempt!)
IDL> print, roundPV( [ 0.0002323,34.3434,1234000], 2)
0.000000 0.000000 1.23400e+006
so I had another think:
function fix_digits, num, digits
p10 = floor(alog10(abs(num)))
expo = 10.0d^(digits -1 - p10)
fix_val = long(num*expo)/expo
return, fix_val
end
this seems to be fine for all thise who dont care about casting to double
- some one else can reset the type!
IDL> print, fix_digits( [ 0.0002323,34.3434,1234000], 2)
0.00023000000 34.000000 1200000.0
Martin
|
|
|
|
| Re: significant figures function? [message #23889 is a reply to message #23884] |
Mon, 26 February 2001 15:44 |
Ben Tupper
Messages: 186
Registered: August 1999
|
Senior Member |
|
|
You try the following.
;------start
;+
; NAME:
; ROUNDPV
;
; PURPOSE:
; This function returns a rounded value to a specified place value for
Integer, Long, Float, And
; Double precision values.
;
; CATEGORY:
; Miscellaneous, Math
;
; CALLING SEQUENCE:
; Result = RoundPV(Value, PlaceValue)
;
;
; INPUTS:
; Value A scalar or vector of type integer, long integer, float or double
precisions.
; PlaceValue An integer specifying the plave value to round toward.
Positive and negative values
; are permitted.
;
; OPTIONAL INPUTS:
; None
;
; KEYWORD PARAMETERS:
; None
;
; OUTPUTS:
; This function returns the rounded value to the spcified place value.
;
; OPTIONAL OUTPUTS:
; None
;
; COMMON BLOCKS:
; None.
;
; SIDE EFFECTS:
; None known.
;
; RESTRICTIONS:
; None known.
;
; EXAMPLE:
; X = 321.489
; For i = -2, 2 Do Print, RoundPV(X, i)
; 321.480
; 321.400
; 321.000
; 320.000
; 300.000
;
; Specifying a negative place value for an integer or long type data value
has no effect. For example
; X = 321L
; For i = -2, 2 Do Print, RoundPV(X, i)
; 321
; 321
; 321
; 320
; 300
;
; MODIFICATION HISTORY:
; Written by: Ben Tupper 30 SEP 1999
; email pemaquidriver@tidewater.net
; tel: (207) 563 - 1048
; 248 Lower Round Pond Road
; POB 106
; Bristol, ME 04539-0106
;
; 6 FEB 2000 Dropped call to home grown TYPE function in favor of SIZE
function
; with STRUCTURE keyword set.
;
;-
FUNCTION RoundPV, X, PV
If N_elements(PV) EQ 0 Then PV = 0
Sz = Size(X, /Str)
Case Sz.Type of
2: X2 = Fix(Float(Long(Float(X)*10.^(-PV)))*10.^PV)
3: X2 = Long(Float(Long(FLoat(X)*10.^(-PV)))*10.^PV)
4: X2 = Float(Long(X*10.^(-PV)))*10.^PV
5: X2 = Double(Long(X*10.d^(-PV)))*10.d^PV
Else:
EndCase
Return, X2
END
;------finish
Med Bennett wrote:
> Has anyone written a function that returns the input value or array with
> a specified number of significant digits? I have lloked at the various
> IDL libraries on the web but did not come up with what I'm after.
> Thanks in advance for any pointers -
--
Ben Tupper
248 Lower Round Pond Road
POB 106
Bristol, ME 04539
Tel: (207) 563-1048
Email: PemaquidRiver@tidewater.net
|
|
|
|
| Re: significant figures function? [message #23891 is a reply to message #23889] |
Mon, 26 February 2001 16:07 |
Martin Downing
Messages: 136
Registered: September 1998
|
Senior Member |
|
|
"Med Bennett" <mbennett@indra.com> wrote in message
news:3A9ACD30.E0F3F5D9@indra.com...
> Has anyone written a function that returns the input value or array with
> a specified number of significant digits? I have lloked at the various
> IDL libraries on the web but did not come up with what I'm after.
> Thanks in advance for any pointers -
>
this should do it
function fix_digits, num, digits
expon = 1.0
while num/expon gt 1 do expon = expon * 10
fix_val = num/expon
fstring = string(digits+2, digits, format = '("(f",i0.0,".",i0.0,")")')
reads, string(fix_val, format = fstring), fix_val
return, fix_val * expon
end
> print, fix_digits(!pi, 3)
3.14000
Martin
----------------------------------------
Mr. Martin Downing,
Clinical Research Physicist,
Orthopaedic RSA Research Centre,
Woodend Hospital,
Aberdeen, AB15 6LS.
m.downing@abdn.ac.uk
Martin
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
