tokenise(string,regexp)
matches(string,regexp)
replace(string,regexp,replacementstring)

tokenise

Returns sequence of substrings that are separated by the matched regular expression

for $w in tokenise(string(.), '\W+') return lower-case($w)
string value of current node is tokenised into words
        (using the \W+ regular expression) and made lower-case
               

matches

Returns true if string matches the regexp

replace

Any string matching regexp in string replaced by replacementstring

replace('banana', 'ana', #)
b#na