Search usernames with Regular Expressions - regex

Question

At the moment I'm trying to use a regular expression to find usernames. The following condition is what I need:
"Username matches the search term with a maximum of 3 wrong characters"
For example,
Database content:
"MyUsername"
Search command -> returning match:
search("Username") -> "MyUsername"
search("Us3rname") -> "MyUsername"
search("userName") -> "MyUsername"
search("MyUser") -> none (4 characters wrong)
search("My Us3r N#me") -> none (4 characters wrong)
I can build my regex dynamically and push this to a database query. I only can't get a grip on the regex itself. Could you help me with this? Would be great? (or is it even possible?)

Answer 1

You can't do this with regular expression. You need some similarity algorithm to check the similarity between two strings.
A good start and an easy one is the levensthein distance.
In short: It calculates how many Insert/Update/Delete Operations are needed to transform string A to string B.
I had done this in Javascript some years ago, but it should be easy in nearly every programming language. You can find a working example here:
// http://ejohn.org/blog/fast-javascript-maxmin/
Array.max = function( array ){
return Math.max.apply( Math, array );
};
Array.min = function( array ){
return Math.min.apply( Math, array );
};
// Levenshstein Distance Calculation
function levenshtein_distance (t1, t2) {
var countI = t1.length+1;
var countJ = t2.length+1;
// build empty 'matrix'
var matrix = new Array (countI);
for (var i=0;i<countI;i++) {
matrix[i] = new Array (countJ);
}
// initialize the matrix;
// set m(0,0) = 0;
// m(0,0<=j<countJ) = j
// m(0<=i<countI, 0) = i
matrix[0][0] = 0;
for (var j=1;j<matrix[0].length;j++) {
matrix[0][j] = j;
}
for (var i=1;i<matrix.length;i++) {
matrix[i][0] = i;
}
// calculate the matrix
for (var i=1;i<matrix.length;i++) {
for (var j=1;j<matrix[i].length;j++) {
var costs = new Array ();
if (t1.charAt(i-1) == t2.charAt(j-1)) {
costs.push (matrix[i-1][j-1]);
}
costs.push (matrix[i-1][j-1] + 1);
costs.push (matrix[i][j-1] + 1);
costs.push (matrix[i-1][j] + 1);
matrix[i][j] = Array.min(costs);
}
}
// resultMatrix = matrix;
var result = new Object
result.distance = matrix[countI-1][countJ-1];
result.matrix = matrix;
return result;
}