generate random numbers in engineering notation with precision

Matlab folks,

I have a simple question: I want to generate a vector of random numbers formatted in engineering format (e-notation) and the precision is x.xxxxExx

The precision is what is causing me trouble especially is that my goal is to generate a vector of these

Thanks

Answer 1

[José-Luis]

By default, variables in Matlab are double. If you want to change the way they are displayed, please look at:

 doc format

If you want to save them to a text file, please look at:

 doc fprintf

[Mike Sirwood]

I know but I lose numbers with such format...for example if I want to generate random between 1e-4 and 1e-8 then I never get numbers like xe-6 or xe-8

[José-Luis]

I don't understand the problem. If you want a format other than double, then you need to rewrite the algorithm.

[Mike Sirwood]

Ok, let me make it simpler, if I used fprintf as you suggested, then how do I store its "formatted output" in a variable as a number?

[José-Luis]

You can't. Not as an "engineering" number. Not unless you create a custom data type. And Matlab would definitely not be the tool I would choose for that job. If I understand correctly your intentions, then you might be better off creating a vector of integer values.

 precision_you_want = 1e-8; %for instance  
 your_interval = 1e-8:precision_you_want:1e-4;
 numRand = 1000;
 your_vector = your_interval(randi(numel(your_interval),1,numRand));

But I still think it is better to round of to the precision you want. There is no "engineering" data type, and mimicking one would be difficult, especially since most computers use binary, not decimal, format.

Answer 2

[Rick Rosson]

http://www.mathworks.com/matlabcentral/fileexchange/39658

[Mike Sirwood]

But when I use this function, for example and using a number like

num2eng(0.00054089)

ans = 540.89e-6

I want it in the format x.xxxxEx and not as a string

Answer 3

[Daniel Shub]

What do you mean by format? Computers tend to work with integers and floats, with floats coming in single and double precision varieties. They are internally represented as a sequence of 0's and 1's and many programs can externally display them in a variety of formats (binary, hex, decimal, engineering, etc). MATLAB by default uses double precision floats, but you can switch to single precision or integers. The format in which the number is displayed is controlled by the FORMAT function, but this has no effect on the internal representation.