Difference between these two lines.

Hi all,

1,

(setq a 26.6)
(setq b (fix a) c (fix (* (abs (- a b)) 10))) => returns [b]6[/b].

2,

(setq a 25.7)
(setq b (fix a) c (fix (* (abs (- a b)) 10))) => returns [b]6[/b].

Why my second point doesn't return me 7 as answer.

Am i doing something wrong..

Please Help.

Weird huh?

(defun _roundTruncated  (a)
     (setq b (fix a)
           c (* (abs (- a b)) 10)
           d (fix (+ c (/ c (abs c) 2.0))))
     )

(_roundtruncated 26.6)

6

(_roundtruncated 25.7)

7

Some variables need to be localized

You can not trust computers these days:

(setq a 25.7)
(rtos (* 10 (- a (fix a))) 2 16)
"6.999999999999992"

So maybe:

(atoi (rtos (* 10 (- a (fix a))) 2 0))
7

You can not trust computers these days:

(setq a 25.7)

[i](rtos (* 10 (- a (fix a))) 2 16)[/i]
[i]"6.999999999999992"[/i]

So maybe:

I believe it has something to do with OS platforms (16/32/64)

The reason why the second code of the OP "fails" is because fix returns a truncated real value. the fix function by definition:

Returns the conversion of a real number into the nearest smaller integer

so (* 10 (- a (fix a))) 2 16) is in fact 6.999999999999992 "fix" to the nearest smallest integer 6

(atoi (rtos (* 10 (- a (fix a))) 2 0))
7

I figured thats what the OP will eventually use for assigning the value to an attribute.

Thanks a LOT pBe, Tharwat, Stefan BMR...

 

I believe it has something to do with OS platforms (16/32/64)

If so, then my old 64bit-p routine may be of use:

http://www.cadtutor.net/forum/showthread.php?65194-lsp-help-for-loading-amp-unloading-arx-file-from-specified-folder!&p=445347&viewfull=1#post445347

HTH

Sounds a lot like the issues in these 2 threads:

http://www.cadtutor.net/forum/showthread.php?60816-round-up-number-in-autolisp

http://www.cadtutor.net/forum/showthread.php?60050-Help-What-wrong-in-Fix-defun

Good show Irneb/Renderman

Stange indeed

 (setq a 25.7)
(setq b (fix a) c (fix (* (abs (- a b)) 100))) 
[b]69[/b]

perhaps

(setq b (fix (* (abs (- a (fix a))) 10.01))) 7

(setq b (fix (* (abs (- a (fix a))) 100.01))) 70

Okay, here's a challenge

(_OneTensHundreds 120) --- "Hundreds"

(_OneTensHundreds 12) ---- "Tens"

(_OneTensHundreds 124.25) --- "Hundreds"

Math only, no conversion to string

or to make it more interesting

(_OneTensHundreds (114.25)

100 10 4 0.20 0.05

Hang on, i think i saw something similar at the swamp. (regarding coin change)

Okay, here's a challenge

 

(_OneTensHundreds 120) --- "Hundreds"

(_OneTensHundreds 12) ---- "Tens"

(_OneTensHundreds 124.25) --- "Hundreds"

 

Math only, no conversion to string

(defun f ( n )
   (cdr (assoc (fix (/ (log n) (log 10))) '((0 . "Units") (1 . "Tens") (2 . "Hundreds") (3 . "Thousands"))))
)

_$ (f 120)
"Hundreds"
_$ (f 12)
"Tens"

Whoa.. that was fast

See updated post Lee. up there /|\

or to make it more interesting

 

(_OneTensHundreds (114.25)

100 10 4 0.20 0.05

Quick hack:

(defun f ( n / x )
   (if (< 0 n)
       (progn
           (setq x (expt 10 (fix (/ (log n) (log 10))))
                 x (* (fix (/ n x)) x)
           )
           (cons x (f (- n x)))
       )
   )
)

_$ (f 123)
(100 20 3)

[ ONLY FOR INTEGERS ]

Quick hack:

 

[ ONLY FOR INTEGERS ]

Neat... I am yet to write my version.. (c'mon google, show me formula....)

Good show Irneb/Renderman

Cheers, Pbe!

Math only, no conversion to string

Translation: "Hey Lee, I bet you can't s o l v e t h i s !?"

(in a taunting voice, which is assured to get Lee's attention)

... Math only

Translation: "Hey Lee, I bet you can't s o l v e t h i s !?"

 

(in a taunting voice, which is assured to get Lee's attention)

 

... Math only

Geez, thats not my intention, but i have to admit it does seem like it. .

Anyhoo. the thing is, with the numeruous functions available for string handling, the challenge would be far too easy for you the REST of us ( oops I did it again ) .

and the solution can be as simple as

(vl-string-position (ascii ".") (rtos n 2))

where's the fun in that? Hence the MATH only

And guys, stop fooling around and start coding

Good show Irneb/Renderman

Okay, here's a challenge

 

(_OneTensHundreds 120) --- "Hundreds"

(_OneTensHundreds 12) ---- "Tens"

(_OneTensHundreds 124.25) --- "Hundreds"

Before your update I was thinking: http://caddons.svn.sourceforge.net/viewvc/caddons/Libraries/Math.LSP?revision=65&view=markup

Functions: int->word (line 1031) and int->wordth (1121)

But those only work on integers, not reals. Mainly due to the problem around this thread. But I think you saw those already from another tread.

Anyhow, I'm thinking about a ~fix routine which adds a minimal fraction before executing normal fix. Might alleviate the problem methinks.

...or to make it more interesting

 

(_OneTensHundreds (114.25)

100 10 4 0.20 0.05

To comply with the challenge criteria:

(defun f ( n / ff )
   (defun ff ( n x / y )
       (if (not (equal n 0.0 1e-)
           (cons
               (setq y (* x (fix (+ (/ n x) 1e-)))
               (ff (- n y) (/ x 10.0))
           )
       )
   )
   (ff n (expt 10.0 (fix (/ (log n) (log 10)))))
)

_$ (f 114.25)
(100.0 10.0 4.0 0.2 0.05)

None too elegant, but dealing with floating point errors is always cumbersome...

To comply with the challenge criteria:

_$ (f 114.25)
(100.0 10.0 4.0 0.2 0.05)

None too elegant, but dealing with floating point errors is always cumbersome...

You should see my code

(defun _breakdown  (num / _dec base a b lst)
     (defun _dec  (l n h b)
           (if (or (< l n h) (= l n))
                 (float l)
                 (_dec h n (+ h b) b))
           )
   [color=blue] [b] (setq base '(10000.00 1000.0 100.0 10.0 1.0 0.1 0.001 0.0001))
[/b][/color]      (while (and (setq a    (car base)
                       base (cdr base))
                 (setq run (_dec 0 num a a)))
           (if (or (/= run 0) (= run a))
                 (setq lst (cons run lst)
                       num (- num (car lst))))
           )
     (reverse lst)
     )

(_breakdown 114.25)

(100.0 10.0 4.0 0.2 0.04);

Also its limited by the base variable.. I may need to add a sub to generate a list depending on the value of n.

Thanks!

Before your update I was thinking: http://caddons.svn.sourceforge.net/viewvc/caddons/Libraries/Math.LSP?revision=65&view=markup

whew , thats a handful. Nice linky

Thanks!

Anyhow, I'm thinking about a ~fix routine which adds a minimal fraction before executing normal fix. Might alleviate the problem methinks.

(setq b (fix (* (abs (- a (fix a))) 10.01))) 7 (setq b (fix (* (abs (- a (fix a))) 100.01))) 70

Hence the challenge Irneb, but it evolves to something else though