vProduct = 1.0e+03 *[
0.0048 0.0076 3.1440
-0.0049 0.0076 3.9840
-0.0017 0.0076 -0.6840
-0.0049 -0.0048 -4.5480
-0.0017 -0.0048 -0.2940
-0.0017 0.0049 -1.3620]
format shorte
vProduct
format longe
vProduct
format bank
vProduct
format shortG
vProduct
It's apparently a new "feature" that the command window displays the "1.00E+00" multiplier even when it is 1; that didn't use to happen. I also think that is annoying and the prior behavior would be preferred.
If one wants a specific format other than the various options given with the format command, one has to write an output format specification and use fprintf which entails counting columns and building the appropriate number of fields and locating the linefeeds. Here's a place where my complaint since the beginnings of MATLAB that the C-style i/o was chosen over Fortran FORMAT really shows up in the inconvenience.
function write2Darray(x,wide,ndec)
% simple-minded function to display array at given precision, width
[r,c]=size(x);
fmt=sprintf('%%%d.%df',wide,ndec); % the single variable format first '%10.3F', e.g
fmt=[repmat(fmt,1,c) '\n']; % repeat to match array width, add newline
fprintf(fmt,x.') % print w/ format; remember transpose
end
write2Darray(vProduct,10,3)
One can get much more elaborate, of course including passing the format type or the width and/or precision by column as vectors.
Before Mathworks made it essentially impossible to add a Fortran compiler I had a mex file that used the Fortran FORMAT facility to do this kind of thing to be able to avoid the repmat construct since FORMAT is defined such that a field repeat multiplier is possible.
ADDENDUM
The problem with the above approach is, of course, one has to manually call it unless bury it inside other functions in which case the semicolon feature doesn't help to turn it off. It's an area I complained about in the most early days that there's no user ability for any finer control over the command output than the ever-existing set of format choices provided.
It seems like I recall a thread a few years ago of a similar complaint/wish and @Walter Roberson discussed some of what would take to override the builtin functionality and it would be, of course, exceedingly difficult if not impossible.
