Rounding Elements of a Matrix to 3 decimal places

Hi all

I'm pulling my hair out with this simple but frustrating issue.

I have two matrices In and Out:-

I am trying to create a new matrix called T which is the subtraction of In and Out.

However i want matrix T to be rounded to 3 dp.

This is what i currently get when i use the round function:-

Can anyone help?

Thank you.

[per isakson]

See format Set Command Window output display format and try

>> format short

[John hope]

Hi Per

thanks for the input, but it still outputs the matrix values to the default of 4 decimal places.

There must be a command that rounds to 3 decimal places???

Thanks

[Star Strider]

You can try this:

x = rand(5, 1);
format long g
y = round(x, 3)

however it doesn’t truly round to 3 decimal places (you can see that easily enough by using format long rather than format long g), although it looks like it does.

Consider using sprintf if this is important.

[Rik]

It looks like 3 decimal places to me.

I admit I mostly use something like this when I need to round to a number of digits:

k=3;
v=1.8247;
v2=10^k*round(v/10^k);

[per isakson]

>> format short g
>> round( y, 3 )
ans =
        0.815        0.632        0.958
        0.906        0.098        0.965
        0.127        0.278        0.158
        0.913        0.547        0.971

[Star Strider]

It is, although only if you let the format function truncate it and ‘round’ it according to the format function’s internal workings.

An emulation of the most recent implementation of round that I came up with a while ago is:

roundn = @(x,n) round(x .* 10.^n)./10.^n;   % Round ‘x’ To ‘n’ Digits, Emulates Latest ‘round’ Function

It produces results identical to the more recent verisons of round that allow rounding to a specific number of decimal places. If you use format long to view the result, it is not actually rounded, only approximated. The format function then truncates it, depending on the desired result.

[John hope]

Thanks Star Strider got it to 3 dp.

Much appreciated.

I must be honest, it seemed awfully complicated to simply output to 3dp but maybe I’m just a little dense.

Thank you.

[Rik]

The visual output of values is indeed easier than the method described here:

val=0.43125;
fprintf('%.3f\n',val)

In Matlab the display of data and changing that data are two unrelated tasks.

[Star Strider]

@John Hope — My pleasure.

If you want engineering notation, use format short eng or format long eng.