[CentralOH] Plucking Digits From Float

[Joe Shaw]

Hi,

Yep, good call.  I wasn't familiar with the Decimal library and hadn't
caught up to your other message on the thread. It definitely looks
like it'll fit the bill.

I would be a bit surprised if the hardware returned floats... most
small embedded components don't have FPUs or IEEE-754 implementations.
 (And for things like medical devices, they'd more likely use some
fixed point implementation to avoid these kinds of rounding errors.)

Joe

On Mon, Dec 9, 2013 at 10:21 AM, Brian Costlow <brian.costlow at gmail.com> wrote:
> There is an existing Decimal library in the standard library, which can be
> used for a lot of this kind of work. It's designed to avoid these kind of
> issues (mostly).
>
> http://docs.python.org/2/library/decimal.html
>
> Although since he is working with hardware, if the boundary can only pass
> floats into Python, he still will have potential issues. But using Decimal
> you can actually trap instances of inexact conversion and deal with them,
> instead of having them sneak by.
>
>
>
>
>
>
> On Mon, Dec 9, 2013 at 9:57 AM, Joe Shaw <joe at joeshaw.org> wrote:
>>
>> Hi,
>>
>> It's worth pointing out also that floating point values are an
>> approximation.  This is one reason why you should never do direct
>> comparisons with floating point numbers, but it could also bite you
>> here.  It's possible that 1.2346 (or any value you choose) cannot be
>> precisely represented as you wish, and so converting it to a string
>> and taking the negative index might be very, very wrong.
>>
>> 0.1 is one such number, and it can be demonstrated in python on some
>> platforms.  Go into the Python REPL and type "0.1" and you might get
>> "0.10000000000000001" back.
>>
>> If this is really important you can store your values as integers --
>> which are exactly precise -- and handle the decimals yourself.  (In
>> other words, your own implemention of fixed point values.)
>>
>> Joe
>>
>> On Sun, Dec 8, 2013 at 2:40 PM,  <jep200404 at columbus.rr.com> wrote:
>> > On Sun, 8 Dec 2013 14:36:01 -0500, Louis Bogdan <looiebwv at gmail.com>
>> > wrote:
>> >
>> >> num is calculated value 1.2346
>> >> a=num(-1) is indicating the numeral "6" with negative indexing
>> >> notation.
>> >> b=num(-2) is indicating the numeral "4"   "         "
>> >
>> > I question the sanity of do so. Nonetheless.
>> >
>> > num = 1.2346
>> > a = str(num)[-1]
>> > b = str(num)[-2]
>> >
>> > print num, a, b
>> >
>> > _______________________________________________
>> > CentralOH mailing list
>> > CentralOH at python.org
>> > https://mail.python.org/mailman/listinfo/centraloh
>> _______________________________________________
>> CentralOH mailing list
>> CentralOH at python.org
>> https://mail.python.org/mailman/listinfo/centraloh
>
>
>
> _______________________________________________
> CentralOH mailing list
> CentralOH at python.org
> https://mail.python.org/mailman/listinfo/centraloh
>