format short and format long

I need to find the relative error between 2 variable, the one in format short and the other in format long. How can I take the same variable, once in short and once in long so I can subtract the two?

[Stephen23]

This question makes no sense. Numeric data is stored in an array of any of the numeric data types (e.g. single, double, uint32, etc), and is displayed in the command window using the currently specified format (e.g. longg, bank, etc). The format has nothing to do with the precision of calculations, nor anything to do with the precision of stored data. Exactly as the documentation clearly states, format is only relevant to the displaying of data, so the statement "I take the same variable, once in short and once in long so I can subtract the two" is totally meaningless.

Answer 1

[Star Strider]

If you have one variable that is single precision and one that is double precision, you can perform your calculation with them as you normally would. The result will be single precision.

According to the documentation on Floating-Point Numbers, specifically Single-Precision Operations:

• You can perform basic arithmetic operations with single and any of the following other classes. The result is always single:

    single

    double

    char

    logical

Answer 2

[John D'Errico]

I think perhaps you misunderstand. Format short does not cause a variable to be stored in any format, nor does format long.

The numbers were always stored in the same precision (usually double precision, unless you specified something else like single.)

So the format command is strictly a tool to change how a number as displayed.