|
View previous topic :: View next topic
|
| Author |
Message |
Soundammal.S
New User
Joined: 05 May 2008
Posts: 29
Location: Chennai
|
|
|
|
Hello ,
I'm working on REXX code in which I have got some values in exponential form. But I need to change it to decimal format.Need help in this.
Thank you. |
|
| Back to top |
|
 |
expat
Global Moderator
Joined: 14 Mar 2007
Posts: 8797
Location: Welsh Wales
|
|
|
|
| What have you researched for yourself and what problems did you encounter ? |
|
| Back to top |
|
|
dbzTHEdinosauer
Global Moderator
Joined: 20 Oct 2006
Posts: 6966
Location: porcelain throne
|
|
|
|
| Soundammal.S wrote: |
| values in exponential form |
I am just a dumb old programmer, could you please explain what you mean by exponential form |
|
| Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006
Posts: 19244
Location: Inside the Matrix
|
|
|
|
Hello,
As with many things, it will help if you post some samples of what you "have" and what you "want as output" when your process runs. . . |
|
| Back to top |
|
|
PeterHolland
Global Moderator
Joined: 27 Oct 2009
Posts: 2481
Location: Netherlands, Amstelveen
|
|
|
|
| Or post "what you have not"? |
|
| Back to top |
|
|
prino
Senior Member
Joined: 07 Feb 2009
Posts: 1306
Location: Vilnius, Lithuania
|
|
|
|
Very simple, use the C2D(value) (He asked for it  ) |
|
| Back to top |
|
|
PeD
Active User
Joined: 26 Nov 2005
Posts: 459
Location: Belgium
|
|
|
|
| Quote: |
| I have got some values in exponential form |
What is the internal representation of an exponential form????
For me, exponential form is 8³ or 7² .... and how can you store that format inside a field ?
Maybe I am wrong !! |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10873
Location: italy
|
|
|
|
wild guess... floating point  |
|
| Back to top |
|
|
wlblaney
New User
Joined: 07 Nov 2009
Posts: 3
Location: Frederick, MD
|
|
|
|
Try using the following routine as a starting point. It will take a floating point number in internal format and translate it itno a human-readable form. You can use the exponent value to determine where to move the decimal point in the mantissa to get a standard decimal number.
| Code: |
xlate_float: procedure
/**********************************************************************/
/*** This routine receives as input a string containing a floating ***/
/*** point number in internal (HPF) format and returns a character ***/
/*** string containing the number in standard notation (0.nnnne99). ***/
/*** The fractional decimal portion of the numeric string returned ***/
/*** will vary depending on the size of the input float variable ***/
/*** and the number of significant digits (that is, right-most non- ***/
/*** zero digit). The value may or may not be negative, and the ***/
/*** exponent may or may not be negative as well. ***/
/**********************************************************************/
parse arg floatin
current_precision = digits()
numeric digits 12
log16 = 1.2041199826559
fracten.0 = 12
fracten.1 = 1.2589254117941
fracten.2 = 1.0232929922807
fracten.3 = 1.0023052380778
fracten.4 = 1.0002302850208
fracten.5 = 1.0000230261160
fracten.6 = 1.0000023025877
fracten.7 = 1.0000002302585
fracten.8 = 1.0000000230258
fracten.9 = 1.0000000023025
fracten.10 = 1.0000000002302
fracten.11 = 1.0000000000230
fracten.12 = 1.0000000000023
floatout = 0
/*******************************************************************/
/*** The internal format of a floating-point number is: ***/
/*** sppp pppp mmmm mmmm mmmm mmmm mmmm mmmm ... ***/
/*** where: ***/
/*** s is the sign bit (0 = positive, 1 = negative); ***/
/*** ppp pppp is the exponent in "excess-64 notation (i.e, ***/
/*** determine the characteristic by subtracting ***/
/*** 64 from the value); ***/
/*** mmmm is 6 or more hex digits that comprise the ***/
/*** mantissa. Increasing the size of a floating ***/
/*** point number increases the accuracy of the ***/
/*** value, but does not increase the range of ***/
/*** possible values. ***/
/*******************************************************************/
/*******************************************************************/
/*** Step 1: Convert the internal format to a hex string: ***/
/*******************************************************************/
hex_string = c2x(floatin)
parse var hex_string exponent 3 mantissa
/*******************************************************************/
/*** Step 2: Get the sign bit: ***/
/*******************************************************************/
sign_bit = substr(x2b(exponent),1,1)
if sign_bit = 0 then,
sign_char = " "
else
sign_char = "-"
/*******************************************************************/
/*** Step 3: Determine the exponent: ***/
/*******************************************************************/
exp16 = x2d(b2x("0"substr(x2b(exponent),2))) - 64
exp10 = exp16 * log16
parse var exp10 expout "." expfrac
if expout < 0 then
expsign = "-"
else
expsign = "+"
/*******************************************************************/
/*** Step 4: Determine the mantissa: ***/
/*******************************************************************/
mantissa = strip(mantissa,'T','0')
mantsize = length(mantissa)
mantissa = x2d(mantissa) / 16 ** mantsize
fracadd = 1
do p = 1 to fracten.0
thispos = substr(expfrac,p,1)
if thispos > 0 then,
fracadd = fracadd * fracten.p ** thispos
end /* do p = 1 to fracten.0 */
if expsign = "+" then,
mantout = mantissa * fracadd
else
mantout = mantissa / fracadd
/*******************************************************************/
/*** Step 5: Step the components together in a readable format: ***/
/*******************************************************************/
numeric digits current_precision
mantout = mantout + 0
floatout = sign_char || mantout * 10 ** expout
return floatout |
Code'd |
|
| Back to top |
|
|
MBabu
Active User
Joined: 03 Aug 2008
Posts: 400
Location: Mumbai
|
|
|
|
| if by 'exponential form' you simply mean the number is showing like 3.14159E12, then see the NUMERIC DIGITS statement |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10873
Location: italy
|
|
|
|
i do not really see why we should waste time for somebody who posts and afterwards,
after having been asked, does not care to answer and provide more info  |
|
| Back to top |
|
|
wlblaney
New User
Joined: 07 Nov 2009
Posts: 3
Location: Frederick, MD
|
|
|
|
Enrico,
While the twit^h^h^h^h colleague who originally posed the question may not deserve any help, there will be many others who might also be interested in a solution. Since I had already solved this very problem, it cost me very little effort to post this. |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10873
Location: italy
|
|
|
|
thanks for the code Bill !
my rant was not specific to this post,
just a general consideration about careless badmannered bahavior |
|
| Back to top |
|
|
prino
Senior Member
Joined: 07 Feb 2009
Posts: 1306
Location: Vilnius, Lithuania
|
|
|
|
Here's some more code that might be useful:
| Code: |
/* This REXX exec computes a S/370 floating-point's value ...
e.g. X'C1280000' represents -2.5
(Michel Castelein, 5 November 1997)*/
say 'Enter the S/370 floating-point (4, 8, or 16 bytes):'
numeric digits 34
pull float
float = X2C(float) /* convert float to a hexadecimal string*/
float_size = LENGTH(float) /* size in Bytes*/
Byte_0 = SUBSTR(float, 1, 1)
select
when BITAND(Byte_0, '80'x) == '00'x then sign = '+'
when BITAND(Byte_0, '80'x) == '80'x then sign = '-'
end
exponent = C2D(BITAND(Byte_0, '7F'x)) - 64
fraction = 0
power = -1
do i = 2 to float_size
if i = 9 then iterate /* skip bits 64-71*/
Next_Byte = C2D(SUBSTR(float, i, 1))
left_Digit = Next_Byte % 16
fraction = fraction + left_Digit * 16**power
right_Digit = Next_Byte // 16
power = power - 1
fraction = fraction + right_Digit * 16**power
power = power - 1
end
interpret 'value =' sign ( fraction * 16 ** exponent )
say "The floating-point's value is" value
exit |
|
|
| Back to top |
|
|
wlblaney
New User
Joined: 07 Nov 2009
Posts: 3
Location: Frederick, MD
|
|
|
|
Robert,
I really like your solution. I didn't think to try using a negative exponent in REXX.
Very elegant. |
|
| Back to top |
|
|
prino
Senior Member
Joined: 07 Feb 2009
Posts: 1306
Location: Vilnius, Lithuania
|
|
|
|
| wlblaney wrote: |
I really like your solution. I didn't think to try using a negative exponent in REXX.
Very elegant. |
It's not mine, but was given to me by Michel Castelein. What I really would like to have is a function that does the inverse, convert a floating point number to an S370 floating point, to finalize my formatted edit function. |
|
| Back to top |
|
|
|
|
