Confusion about precision

Question

When I evaluate

x = 1.0000000000000000001 (* Precision of 20 digits *)

Mathematica returns

1.000000000000000000 (* Precision of 19 digits *)

When I evaluate

y = N[1.0000000000000000001,100]

it still returns

1.000000000000000000

despite the fact that we did know the original number with a precision of 20 digits so it seems it really does lose precision the moment it evaluates the number 1.0000000000000000001. Trace also shows the calculation as if I just entered the number with a precision of 19 digits When evaluate

SetPrecision[y,30]

it returns

1.00000000000000000010000000000

though...

I read through the documentation on arbitrary precision calculations and still don't get what's going on. If Mathematica stores the numbers internally with a higher precision then why doesn't N give me the result up to the highest precision the number has. If it doesn't store any precision then why does the missing digit suddenly appear when evaluating SetPrecision?

Note that InputForm and RealDigits both indicate that last digit is present. In[14]:= InputForm[x = 1.0000000000000000001] RealDigits[x] Out[15]= {{1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1}, 1} 1.000000000000000000119.` — Daniel Lichtblau, Apr 05 '22 at 13:54

Learning Lin · Answer 1 · 2022-04-05T11:54:09.440

2

1.0000000000000000001
1.000000000000000000

It's precision of display....

Have you learned C or C++? It's like:

printf("%.3d",1.000001);

If you mean:

1.0000000000000000001==1.0000000000000000000

It's because the default precision has only ten numbers, You need usesetPrecision or like

1.0000000000000000001`22==1.0000000000000000000`22

There is refered in Possible Issues (https://reference.wolfram.com/language/ref/Equal.html)

It's mean setPrecision[y,30] make Precision[y] from 10 to 30... In other way, the infinite precision is you can set some value's precision infinite, not any of them is infinite defaultly.

(If you think my English is poor, it's because I'm not native English speaker....

edited Apr 05 '22 at 11:54

answered Apr 05 '22 at 11:18

Learning Lin

80
5

But then why does Round[1.0000000000000000001] ==1.000000000000000001 return True? – Gert Apr 05 '22 at 11:20
@Gert why not? Round[1.0000000000000000001] ==1==1.000000000000000000 – Learning Lin Apr 05 '22 at 11:21
sorry, I copied the wrong number. Edited my comment. – Gert Apr 05 '22 at 11:23
Emm, I need to try it in my computer.... – Learning Lin Apr 05 '22 at 11:26
@Gert The "==" in documentation said, in my opinion, the default precision of value is automatic precision ,it only care about the first ten numbers in 1.0000000000001,so.... – Learning Lin Apr 05 '22 at 11:35
@Gert you need use like 1.0000000000000000001`22==1.0000000000000000000`22 – Learning Lin Apr 05 '22 at 11:37
@Gert Is what you want to know? I mean do you understand? My English is poor... so if you are puzzled, please tell me.... – Learning Lin Apr 05 '22 at 11:57

Confusion about precision

1 Answers1