Nothing, that would probably be my preferred choice too (Ric pointed
this out subsequently).
I was just focusing on the real question of print precision - the same
print precision issue occurs for either approach.
Note though that (- <.) performs faster albeit uses more space.
dec1=:- <.
dec2=:1&|
10{.z=:0.01*?100000$100000
318.72 415.11 155.08 826.2 664.99 381.41 148.83 758.18 562.25 117.54
ts=: 6!:2 , 7!:2...@]
100 ts 'dec1 z'
0.00573867950439 1573696
100 ts 'dec2 z'
0.010124067688 1049472
(dec1 -: dec2) z
1
Thanks Oleg, Regards Rob
On 03/05/2009, at 8:35 AM, Oleg Kobchenko wrote:
>
> What's wrong with (1&|) ?
>
>
> ((- <.) -: 1&|) _5.1 _5.01 _0.0001 0 0.0001 5.1 5.2 5.0002
> 1
>
>
>> From: Rob Hodgkinson <[email protected]>
>>
>> Alex, you may be asking 'too much print precision' on your display,
>> eg:
>>
>> decimal=:-<.
>> 9!:11 ]14
>>
>> decimal 5.1 5.2
>> 0.1 0.2
>> 9!:11 ]15
>>
>> decimal 5.1 5.2
>> 0.0999999999999996 0.2
>> 9!:11 ]16
>>
>> decimal 5.1 5.2
>> 0.09999999999999964 0.2000000000000002
>>
>> You may need to use fmt (8!: x) to produce a specified format to
>> avoid
>> the system rounding display.
>>
>> Rob Hodgkinson
>>
>>
>> ----------------------------------------------------------------------
>> For information about J forums see http://www.jsoftware.com/
>> forums.htm
>
>
>
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
