equality operator between matrix and scalar

consider:

a=[-1:0.1:1];
c=a==.1;

it returns c as a matrix of nulls; while I expect c(12) to be 1.

Answer 1

[Matt Kindig]

This is due to floating point precision errors, explained here: http://blogs.mathworks.com/loren/2006/08/23/a-glimpse-into-floating-point-accuracy/

A better way to do this is by comparing against a tolerance, such as:

c = abs(a-0.1)<=eps

Answer 2

[Kye Taylor]

The value .1 and a(12) differ by only one bit as a result of numerical round-off. You can see this with the commands

num2hex(.1)
num2hex(a(12))

A better way to test equality takes into account the possibility of numerical round-off. For example, create your new c variable with the command

c = abs(a-0.1)<=eps(max(a));

Answer 3

[Walter Roberson]

http://matlab.wikia.com/wiki/FAQ#Why_is_0.3_-_0.2_-_0.1_.28or_similar.29_not_equal_to_zero.3F

Answer 4

[James Tursa]

You can try this to see what the numbers are exactly:

num2strexact((-1:0.1:1)')

Sometimes a beter way to do this is to have an array with integer values to start with and then create a 2nd array with the fractional values. E.g.,

A = -10:10;
a = A / 10;
c = A==1;

You can use a for downstream calculations just like before, and you can use c for the value testing.

You can find num2strexact on the FEX here:

http://www.mathworks.com/matlabcentral/fileexchange/22239-num2strexact-exact-version-of-num2str