numbers to the left of the dot in tab of matlab

Hi everyone, I'm new and I'm not sure how this format works, I don't know if I asked the question correctly but I hope you can help me. I have to calculate iter error sol res in the matlab tab with decimal and integer digits, in fixed and floating point format. how can i know how many digits to the left of the dot I have to put?

example

fprintf('%3d %24.5e %12.2e\n' ,tab')

Answer 1

[Walter Roberson]

The %e format always outputs exactly one digit before the decimal place, and outputs a number of decimal places after that according to the format you give, such as 5 decimal places for the %24.5e format. So far that is 1 leading + 5 after decimal [in this case] + decimal place itself, subtotal 7 so far. It always always has a sign character for the exponent, so subtotal 8 now.

If the value is negative, then it will use one place for that. If your leading width such as the 24 in %24.5e is not wide enough to accomodate everything including the leading - then it will use an extra character .

Now we get to the question of how many digits of exponent there are. If I recall correctly, the rule is:

• if you are on MS Windows, or abs() of the exponent is 100 or greater, then 3 digits of exponent are used
• if you are on Mac or Linux and abs() of the exponent is 99 or less, then 2 digits of exponent are used

So take the subtotal of 8, add 3 potential digits for exponent, add 1 potential place for sign, for a total of 12 in this case (7 plus number of digits you specified after the period in the %.e format). No value will require more than that.

[Chiara]

In% 24.5e how do I know what to add from subtotal 8 to 24?

and for these other two examples which methods should I use?

1) fprintf('%3d %11.3e %11.3e\n',tab')

2) printf('%2d %8.5f %8.5f %8.5f %8.5f %8.5f %8.5f %9.1e %9.1e\n',tab')

[Walter Roberson]

In %N.5e you will get one of the following

123456789ABCD
x.xxxxxe+xx      %positive number, small exponent, linux or mac
x.xxxxxe+xxx     %positive number, large exponent, or any time on pc
-x.xxxxxe+xx     %negative number, small exponent, linux or mac
-x.xxxxxe+xxx    %negative number, large exponent, or any time on pc

That could be 11, 12, or 13 characters depending on the value range and the platform.

Now, for %N.5e if N > number of characters required according to the above, then the whole thing will be written at the right hand side of N character positions, so N minus however many blanks proceeding it -- 13, 12, or 11 leading blanks according to the above.

If N < number of characters required according to the above, then N will effectively be ignored, and all of the characters defined above will be inserted.

%Nd is an integer filling the right hand side of a mininum of N spaces. If the input is not an integer, a floating point number will be output, messing up the formatting.

%N.5f uses exactly 5 digits after the decimal point, plus the decimal point, plus a minimum of one leading digit, plus a leading negative sign if required, for a minimum of 7 characters, such as

1234567
0.31415

If the value is 10 or greater, more digits will be used to the left. The number will not be converted into exponential format -- so for example pi*10^50 would output 50 digits before the decimal place (many of them zero.) Again, the value will be placed on the right hand side of a field of blanks of minimum width N.

[Chiara]

Thank you!