Print number as .0000000E+00

I would like to print a number with the following format:

.0000000E+00

MATLAB prefers the following:

0.0000000E+000

Is there a way to solve this issue?

Answer 1

[Jan]

Java offers more methods to influence the output. See:

• http://undocumentedmatlab.com/blog/formatting-numbers/
• http://docs.oracle.com/javase/6/docs/api/java/text/DecimalFormat.html
• http://blogs.mathworks.com/community/2009/07/06/calling-java-from-matlab/

[Hamilton]

This is helpful removing the leading zero. Is there a way to maintain the exponential format? I think i can take care of the E+000 to E+00 issue by simply using some string modification functions.

[Jan]

Perhaps something like this:

char(java.text.DecimalFormat('#,##0.###E00').format(0.1));

Answer 2

[Pourya Alinezhad]

fprintf('%1.5s',0.1)
1.5 stands for 5 decimal places.1 is for full number.
s==scientific
you could easily find a way to this problem if you were a c proogrammer...

[Jan]

No, according to the documentation '%s' means 'string', not 'scientific'. I assume you mean '%e' for the scientific notation.

But to my surprise the output is exactly the same (tested under 2009a, 2011b). This means that '%s' has the undocumented feature to act like '%e' when it gets numbers as input. Strange.

Anyhow, Hamilton asks for an omitted leading zero and a 2 digit exponent. While the first is ugly and provokes unexpected errors, such that it will hopefully not implemented, the second is the default in non-PC versions of Matlab and in the newest release R2013a.

[Pourya Alinezhad]

actually i didn't checked this code in matlab.i used to code in this way in turbo c.and i per-assumed that this is true in matlab :)

[Jan]

@Pourya: And obviously your assumption is not really wrong. Thanks for this interesting new piece of information.

Of course TMW did not implement fprintf from scratch but relies on existing libraries. Therefore even known bugs like buffer overflows and other undocumented features like %s with numbers appear in Matlab.