MeasureThat.net

Run results for: Is it worth checking for empty string before replace with regex?

Go to the benchmark
Run details:
Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:136.0) Gecko/20100101 Firefox/136.0
Firefox 136
Mac OS X 10.15
Desktop
12 days ago
Test name Executions per second
trim and always replace 2453.7 Ops/sec
trim and conditionally replace 2604.8 Ops/sec
trim and conditionally replace with if 2607.7 Ops/sec
always replace and trim 2255.7 Ops/sec
conditionally trim and conditionally replace 2641.4 Ops/sec
Script Preparation code:
​x
 
const arr = []
for(let i = 1; i < 1000; i++) {
    arr.push(' '.repeat(i))
}
for(let i = 1; i < 100; i++) {
    arr.push('');
    arr.push(' text '.repeat(i))
}
​
const run = (fn) => {
    const result = []
    for(const val of arr) {
        result.push(fn(val));
    }
    return result
}
Tests:

  • trim and always replace

     
    run((val) => val.trim().replace(/\s+/g, ' '))

  • trim and conditionally replace

     
    run((val) => {
      const trimmed = val.trim();
      return trimmed && trimmed.replace(/\s+/g, ' ');
    })

  • trim and conditionally replace with if

     
    run((val) => {
      const trimmed = val.trim();
      if (trimmed === '') {
        return '';
      } else {
        return trimmed.replace(/\s+/g, ' ');
      }
    })

  • always replace and trim

     
    run((val) => val.replace(/\s+/g, ' ').trim())

  • conditionally trim and conditionally replace

     
    run((val) => {
      if(val === '') {
        return val;
      }
      const trimmed = val.trim();
      if (trimmed === '') {
        return trimmed;
      } else {
        return trimmed.replace(/\s+/g, ' ');
      }
    })