Bug in max function?

The max function in Matlab gives unexpected results.

When I type

max(double(1000), uint8(1))

matlab gives the result

ans = 255

It looks like the double is silently converted to an uint8, which in that case would have the value 255, and is thus returned as the maximum. It seems more sensible to me to automatically convert all numbers to the largest number class.

Is this a bug or is this behaviour documented somewhere?

Answer 1

[James Tursa]

Documented behavior. When combing double with lesser classes (single, uint8, int32, etc) the conversion is done to the lesser class. This has been discussed before in this forum.

[Daniël]

Thanks for the quick answer.

However, this feature has caused me to spend some time figuring out why the following was wrong:

image = imread(file_name);

histo = hist(image(:), linspace(0, 360, N));

with the error:

??? Error using ==> histc

Edge vector must me monotonically non-decreasing.

Error in ==> hist at 92