So, I've written (for lack of answers + improving my skills) a search script that would do basically what .indexOf does.
function search(ref, data) {
var x
var y
var result = []
if (data == '' || data == null) {
} else {
for (x = 0; x < data.length; x++) {
if (data[x] == ref[0]) { //achando match inicial
var proto = [];
for (y = 0; y < ref.length; y++) {
if (data[x+y] == ref[y]) { //gravando tentativas de match completo
proto.push(data[x+y])
}
}
var proto2 = proto.join('')
if (proto2 == ref) { //testando match completo
result.push(x)
}
}
}
}
if (result == '' || result == null) {
} else {
return result[0]
}
}
It works fine within other little codes and custom functions that do not require too much looping, but when I wrote a more robust script I found that my code is roughly 3000x slower than the native .indeOf.
Why would I incur in such a difference?
Your function is comparing each character from ref
separately with a data
character, in an inner loop, pushing each character match inside an array (proto
) and using join
to transform the array back to a string (proto2
), before comparing it to the original ref
.
This is extremely inefficient and could be greatly simplified, so no question this function is much slower than String.prototype.indexOf().
You could have a single loop instead, iterating through data
, and for each iteration, retrieve a slice of data
with the same length as ref
, before comparing this slice with ref
. If both values match, the string ref
has been found in data
, and the corresponding index is returned:
function searchString(ref, data) {
for (let i = 0; i <= data.length - ref.length; i++) {
if (ref === data.slice(i, i + ref.length)) return i;
}
return -1;
}
While using this function, I get similar execution times than with indexOf
.