How to read a very large number, say of 700 digits?

If I have a number, say N of 700 digits, then how can I read it?

I have tried using ELLIPSIS (...) as:

N = 1234567890123 ...

4567899089339 ...

1234567891233;

But, this does Not work.

[Walter Roberson]

What datatype are you hoping the result will be ?

[Abdul Gaffar]

a number N, like N = 1234... (up to 700 digits). i.e., N is a 1x1 double format.

Answer 1

[John D'Errico]

You don't understand. A 700 digit number is not representable in double precision. Even if it were, then roughly 684 of those digits will never be stored in the number when you do put it into a variable. Anything past the 16 most significant digits will be complete garbage. For example:

X = 1234567890987654321012345;

X has 25 digits in it.

format long g
X

X = 
      1.23456789098765e+24

But have we stored al of those digits, correctly?

sprintf('%0.30f',X)

ans = '1234567890987654441861120.000000000000000000000000000000'

MATLAB sees it as an integer, but do you see that the 9 least significant digits are complete garbage?

Next, can you use a line continuation for the digits of a number? No. So you CANNOT do this to represent a 20 digit number

Y = 1234567890...

9876543210;

MATLAB will generate an error if you try.

The only way you could store a number like that is if you use symbolic form. Thus...

S = '123456789012345678901234567890123456789012345678901234567890'

S = '123456789012345678901234567890123456789012345678901234567890'

XS = str2sym(S)

XS = 

123456789012345678901234567890123456789012345678901234567890

And you could build up a string vector using multiple continued lines. But don't bother trying to put that number into a double. Again, complete crapola.

But remember that all computations with such a number will be incredibly slow, when compared to using double precision. You could also use my VPI or HPF toolboxes to represent long numbers like this, but again, quite slow by comparison.

[Stephen23]

Emphasizing: 700 digits is well above the largest floating point number anyway.

realmax()

ans = 1.7977e+308

Answer 2

[Sulaymon Eshkabilov]

A good solution for this exercise is vpi() fcn developed by John D'Errico:

https://www.mathworks.com/matlabcentral/fileexchange/22725-variable-precision-integer-arithmetic?s_tid=mlc_recom_leaf

Answer 3

[Steven Lord]

Your 700 digit number is too large to fit in double precision. The largest finite double precision number doesn't have that many digits.

log10(realmax)

ans = 308.2547

What were you hoping to do with this huge number? Are you hoping to deploy whatever solution you come up with using MATLAB Compiler or MATLAB Coder?

[Abdul Gaffar]

Maybe I don't undertand you. I am rewriting:

Let N = 123456...23, (a 700 digits number) of double type.

One way to read N is as: Write N in a single row, and I am done.

But, if I have to write N in several rows, as I write the following (using ellipsis):

y = 3 + ...

2 + ...

3

then how to do for N?

Further, I will separate every 2 digits of N to make N as 1 x 350 row vector. If more elaboration is needed, I am ready to provide.

[Stephen23]

"Let N = 123456...23, (a 700 digits number) of double type."

Nope. As this answer already explains, that value is too large to store as DOUBLE type.