While working on the reporting system, I needed to pass a test by comparing .xlsx spreadsheets built with our system with those created with BIRT. Reports were compared by scripts, row-by-row. I identified and resolved all apparent differences, but the test was still failed.

Cells in both files visually looked the same. There were numbers with spaces as decimal separators.

Are they different?

I decided to copy its values to IDE and compare them by Ctrl + F. Search had to show two matches, but the match was only one. Cell values were not the same! I started comparing strings one symbol after another and found that the difference was in spaces.

Quick test written in JS showed the difference:

1
2
3
4
5
'58 971 713'.charCodeAt(2);
// > 160

'58 971 713'.charCodeAt(2);
// > 32

BIRT used non-breaking space as the decimal separator (and it was correct), and I used the common space in my reporting system.

So, there are two types of space characters - Space and Non-breaking space. They look identical, but they are the different symbols, and they are not equal. Non-breaking space is used to prevent automatic line breaking.

I didn’t know about it because I haven’t worked with typography a lot. After changing the separator symbol, all tests became green.

Conclusion

Different symbols can look the same (for example: latin c and cyrillic с). And space is one of these symbols, too. It can be just a space, or non-breaking space.

Even if these symbols look similar, they are different for computers and software. This detail, invisible for a human eye, can be the reason of an error.

BTW, there are more types of spaces, but they don’t look like the common space. Here they are: