Potential bug in pdist or variable savings??

Hi,

I've been running a little program which generates a vector called "past_ctry". Everything in past_ctry is evaluated as 48.281. However when I use this variable and explicit 48.281 in pdist. I got different result with one is 0.001005 and the other is 0.00074012. Could someone explain why? The data can be found in this link:

https://www.dropbox.com/s/lfysfecrjwbqc7n/abc.mat?dl=0

And the program is following:

clear all; close all; clc;
load('abc')

x1 = 1.573;
y1 = 48.282;
x2 = 1.5731;
y2 = 48.281;
y2_new = past_ctry(1); % past_ctry(1) is also 48.281

pdist([x1, y1; x2, y2]) % result is 0.001005
pdist([x1, y1; x2, y2_new]) % result is 0.00074012

Answer 1

[John D'Errico]

Congratulations! You are the 10 millionth person to find a this "bug".

Except, it is not a bug. What has happened is the actual value of one (or both) of those numbers is not actually 48.281. It is probably just a wee bit off, and it was displayed using limited precision. So you THINK the value is exactly 48.281, but, NOT.

Watch out when a number is displayed with only 5 digits of precision. Numbers are stored as doubles in Matlab.

[Walter Roberson]

What is stored is

48.28126666666667432536996784619987010955810546875

which is about 0.000266666666675519 larger than 48.281