Merge dni into master

README.md

@@ -52,6 +52,14 @@ Scan a file, like for example a sitemap.xml file, for urls:
 
 This feature is useful to create files with urls to use with [check-my-pages](https://github.com/greenpeace/check-my-pages).
 
+#### Scan Spanish ID numbers
+
+Scan Spanish ID numbers (DNI or NIE)
+
+```bash
+./ecounter -input=rawfile.csv -count=dnis -output=uniques.txt
+```
+
 #### Hash emails (encrypt)
 
 ```bash

ecounter.go

@@ -11,6 +11,8 @@ const emailRegex string = `([a-zA-Z0-9_\-\.]+)@((\[[0-9]{1,3}\.[0-9]{1,3}\.[0-9]
 
 const shaRegex string = `[A-Fa-f0-9]{64}`
 
+const dninieRegex string = `[A-z]?\d{7,8}[TRWAGMYFPDXBNJZSQVHLCKEtrwagmyfpdxbnjzsqvhlcke]`
+
 const urlsRegex string = `https?://([\da-z\.-]+)\.([a-z\.]{2,6})([/\w \.-]*)*/?`
 
 var debug *bool
@@ -37,6 +39,7 @@ func main() {
 
 		var allMatches []string
 		var allMatchesLC []string
+		var allMatchesUC []string
 		var uniques []string
 
 		switch *countIt {
@@ -51,6 +54,10 @@ func main() {
 		case "urls":
 			allMatches = searchInString(fileToHandle, urlsRegex)
 			uniques = uniquesInArray(allMatches)
+		case "dnis":
+			allMatches = searchInString(fileToHandle, dninieRegex)
+			allMatchesUC = arrayToUpercase(allMatches)
+			uniques = uniquesInArray(allMatchesUC)
 		default:
 			allMatches = searchInString(fileToHandle, emailRegex)
 			allMatchesLC = arrayToLowercase(allMatches)

functions.go

@@ -56,6 +56,17 @@ func arrayToLowercase(a []string) []string {
 	return result
 }
 
+// arrayToUpercase Upercases all the items in an array.
+func arrayToUpercase(a []string) []string {
+	defer timeTrack(time.Now(), "arrayToUpercase")
+	var result []string
+	for _, v := range a {
+		lv := strings.ToUpper(v)
+		result = append(result, lv)
+	}
+	return result
+}
+
 //uniquesInArray Finds uniques in an array and returns.
 func uniquesInArray(a []string) []string {
 	defer timeTrack(time.Now(), "uniquesInArray")

help.go

@@ -16,7 +16,7 @@ Use the options as in this example:
 
 -help				Display this help
 -input=rawfile.csv		Define the input file as rawfile.csv
--count=emails			What to count in the file. It can be "emails", "sha256" or "urls". By default it counts emails.
+-count=emails			What to count in the file. It can be "emails", "sha256", "urls" or "dnis". By default it counts emails.
 -output=uniques.txt		Define the output file as uniques.txt 
 -encrypt=true			Encrypt the results as sha256, to upload to Google Adwords
 -debug=true			Debug the script