Light Regular Expressions

matches and replaceAll cover most of what a tester needs from regular expressions — validating an id's shape, or stripping everything that isn't a digit.

1 min read

A regular expression is a pattern language for text. You do not need to master it, and for test automation a small vocabulary covers most needs. matches asks whether a whole String fits a pattern:

String[] inputs = {"DEF-101", "def-101", "DEFECT-101", "DEF-9999"};
for (String in : inputs) {
    boolean ok = in.matches("DEF-\\d{3}");
    System.out.println(in + " -> " + ok);
}

DEF-101 -> true
def-101 -> false
DEFECT-101 -> false
DEF-9999 -> false

The pattern DEF-\d{3} reads as: the literal text DEF-, then exactly three digits. \d means a digit, and {3} means exactly three of them. In Java source you write \\d because the backslash itself must be escaped. All three rejections are correct: wrong case, wrong prefix, four digits instead of three.

replaceAll uses the same pattern language to strip or transform:

String text = "Order ORD-4471 shipped";
String digits = text.replaceAll("[^0-9]", "");

digits only: 4471

[^0-9] means "any character that is not a digit," so replacing all of them with nothing leaves the number. That one line handles a great many scraping problems.

A word of restraint: regular expressions are powerful and become unreadable quickly. For validating an id format or pulling digits out of a label they are ideal. For parsing something with real structure, such as JSON or HTML, use a proper parser instead. A regular expression will appear to work, then fail on the first input you did not imagine.

Splitting and pattern-matching both assume the text is well-formed enough to process. Real test data is not always that cooperative — sometimes the cell is blank, or holds "n/a" instead of a number.