Regex collating symbols

Question

I tried to understand how 'collating symbols' match works but I did not come out this. I understood that it means matching an exact sequence instead of just the character(s), that is:

echo "ciiiao" | grep '[oa]'     --> output 'ciiiao'
echo "ciiiao" | grep '[[.oa.]]' --> no output
echo "ciiiao" | grep '[[.ia.]]' --> output 'ciiiao'

However, the third command does not work. Am I wrong or I misinterpret something?

I have read this regexp tutorial.

Answer 1

Collating symbols are typically used when a digraph is treated like a single character in a language. They are an element of the POSIX regular expression specification, and are not widely supported.

For example, the Welsh alphabet has a number of digraphs that are treated as a single letter (marked with a * below)

a b c ch d dd e f ff g ng h i j l ll m n o p ph r rh s t th u w y
       *           *    *          *          *    *      *

Assuming the locale file defines it (a collating symbol will only work if it is defined in the current locale), the collating symbol [[.ng.]] is treated like a single character. Likewise, a single character expression like . or [^a] will also match "ff" or "th." This also affects sorting, so that [p-t] will include the digraphs "ph" and "rh" in addition to the expected single letters.