Golang : Normalize unicode strings for comparison purpose

Here is a tutorial on of how to normalize unicode strings and do proper comparison of unicode strings. Suggest that you at least read https://blog.golang.org/normalization once before reading this tutorial further to have a better grasp on the problem.

Normalization of unicode strings plays an important role in ensuring your program will process the user input properly. It is also useful in performing sanity checks on incoming data to ensure the underlying representation matched before performing further strings manipulation. Having different representations can cause your program to produce inaccurate strings comparison result.

In this example, comparing two non-normalized unicode strings will result in mis-match. Not because of their length difference, but because of the strings' underlying representation does not match each other. Normalizing the unicode strings with transform.Chain() functions will create new strings with matching underlying representation.

 package main

 import (
 "fmt"
 "golang.org/x/text/transform"
 "golang.org/x/text/unicode/norm"
 "strings"
 "unicode"
 )

 func isMn(r rune) bool {
 return unicode.Is(unicode.Mn, r) // Mn: nonspacing marks
 }


 func main() {
 str1 := "ElNi\u00f1o"
 str2 := "ElNin\u0303o"

 fmt.Printf("%s length is %d \n", str1, len(str1))
 fmt.Printf("%s length is %d \n", str2, len(str2))

 match := strings.EqualFold(str1, str2)
 fmt.Println(match)

 fmt.Println("Normalizing unicode strings....")

 t := transform.Chain(norm.NFD, transform.RemoveFunc(isMn), norm.NFC)
 normStr1, _, _ := transform.String(t, str1)
 fmt.Printf("%s length is %d \n", normStr1, len(str1))

 normStr2, _, _ := transform.String(t, str2)
 fmt.Printf("%s length is %d \n", normStr2, len(str2))

 match2 := strings.EqualFold(normStr1, normStr2)
 fmt.Println(match2)
 }

Output:

ElNiño length is 7

ElNiño length is 8

false

Normalizing unicode strings....

ElNino length is 7

ElNino length is 8

true

Happy coding!

NOTES: Why the length is not the same even after the strings are normalized? Because \u00f1o is composed of a ñ character and \u0303o is composed of n and ~ characters.

References:

https://socketloop.com/tutorials/golang-strings-comparison

https://blog.golang.org/normalization

http://stackoverflow.com/questions/26722450/remove-diacritics-using-go

https://godoc.org/golang.org/x/text/unicode/norm