I'm trying to remove all characters from a String, except

• numbers
• letters (ignoring case)
• comma (,) and
• dashes (-)

Actually I built some regex that works:

.replaceAll("\\[", "").replaceAll("\\]", "").replaceAll("\"", "")

Example

Given that input with UUIDs formatted as string:

["edfcb406-b37e-4729-899e-93ea2af71e83","98e75d74-abe2-4c08-8340-06aa9b2faf0b"]

I want as expected result:

edfcb406-b37e-4729-899e-93ea2af71e83,98e75d74-abe2-4c08-8340-06aa9b2faf0b

Is there a better regex to get the expected result?

CodePudding user response：

Use Wiktor's commented solution.

Regular expression to clean UUIDs

A regex to clean UUIDs:

[^-,a-zA-Z0-9]

It replaces all non-UUID characters (except separating commas). The regex matches using a negated (^) range of allowed UUID characters:

• dashes -
• commas , (to allow a comma-separated list of UUIDs)
• 26 lowercase a-z and uppercase A-Z alphabetic letters
• digits 0-9

If they are excluded from replacement with empty-string, then all other (dirty / noisy characters) are removed. In fact, this cleans to result in a comma-separated UUID list (CSV).

See the demo on IDEone:

String input = "[\"edfcb406-b37e-4729-899e-93ea2af71e83\",\"98e75d74-abe2-4c08-8340-06aa9b2faf0b\"]";
System.out.println("Input: "   input);
String ouputAttempt = input.replaceAll("\\[", "").replaceAll("\\]", "").replaceAll("\"", "");
System.out.println("ouputAttempt: "   ouputAttempt);
String cleanedUUIDs = input.replaceAll("[^-,a-zA-Z0-9]", "");
System.out.println("cleanedUUIDs: "   cleanedUUIDs);

Prints expected:

Input: ["edfcb406-b37e-4729-899e-93ea2af71e83","98e75d74-abe2-4c08-8340-06aa9b2faf0b"]
ouputAttempt: edfcb406-b37e-4729-899e-93ea2af71e83,98e75d74-abe2-4c08-8340-06aa9b2faf0b
cleanedUUIDs: edfcb406-b37e-4729-899e-93ea2af71e83,98e75d74-abe2-4c08-8340-06aa9b2faf0b

Bonus: comma-separated values to a collection

Then you can split the result string (by comma) to get

• an String[] array of UUIDs or
• a List (both may contain duplicates) or
• a Set (for only unique values)

like:

String[] arrayOfUUIDs = cleanedUUIDs.split(",");
List<String> sequenceOfUUIDs = Arrays.asList(arrayOfUUIDs); // ordered, but not unique (may contain duplicates)
Set<String> uniqueUUIDs = Set.of(arrayOfUUIDs); // unique, but not ordered

See also:

• Searching for UUIDs in text with regex
• iHateRegex: Regex for uuid
• Regex Pattern: UUID & GUID Regular Expression
• Java docs: class java.util.UUID, and its factory-method UUID.fromString(s) to parse a regex from String

CodePudding user response：

A great idea from xehpuk's comment: Looks like a JSON string of UUIDs.

Parsing UUIDs given as JSON array of strings

Your given input ["edfcb406-b37e-4729-899e-93ea2af71e83", "98e75d74-abe2-4c08-8340-06aa9b2faf0b"], coincidentally (or not), is valid JSON (try it), an array containing 2 strings (UUIDs).

JSON applies

A JSON array basically is an ordered list of zero or more elements (e.g. quoted strings) enclosed in square-brackets, like this: Either [] empty or filled ["A", "", null, " 0815", "Long-text or number 7, even quotes escaped\" allowed."]. Both are valid JSON an can be parsed independent from platform & programming-language. The former is an empty list, the latter is a representation of these 5 strings:

• A
• empty string
• null as null-reference in Java
• 0815 number
• Long-text or number 7, even quotes escaped" allowed. (with double-quote inside)

Using Jackson to parse/extract UUIDs from JSON-format

So we can parse that using a JSON parser in Java, like FasterXML's Jackson.

Adapted and extended from:

• Map JSON string array to List<String> using Jackson

import java.io.IOException; // can be thrown when reading/writing JSON

// basic Jackson classes for parsing JSON 
import com.fasterxml.jackson.core.JsonParseException;
import com.fasterxml.jackson.databind.ObjectMapper;
import com.fasterxml.jackson.databind.type.TypeFactory;

String json = "[\"edfcb406-b37e-4729-899e-93ea2af71e83\",\"98e75d74-abe2-4c08-8340-06aa9b2faf0b\"]";
ObjectMapper mapper = new ObjectMapper();

List<String> list = mapper.readValue(json, TypeFactory.defaultInstance().constructCollectionType(List.class, String.class));

String[] array = objectMapper.readValue(json, String[].class);
List<String> anotherList = objectMapper.readValue(json, new TypeReference<List<String>>(){});

See also:

• Convert JSON Array to a Java Array or List with Jackson
• Baeldung: Jackson - Marshall String to JsonNode, tutorial on the low-level JsonNode tree that can be traversed to get all values (e.g. text-values inside an array)
• Baeldung: articles tagged with Jackson