I have already parsed a string up to index idx. My next parse step uses a Regexp. It needs to match the next part of the string, i.e. staring from position idx . How do I do this efficiently?

For example:

let myString = "<p>ONE</p><p>TWO</p>"
let idx

// some code not shown here parses the first paragraph
// and updates idx
idx = 10

// next parse step must continue from idx 
let myRegex = /<p>[^<]*<\/p>/
let subbed = myString.substring(idx)
let result = myRegex.exec(subbed)
console.log(result) // "<p>TWO</p>", not "<p>ONE</p>"

I have already parsed a string up to index idx. My next parse step uses a Regexp. It needs to match the next part of the string, i.e. staring from position idx . How do I do this efficiently?

For example:

let myString = "<p>ONE</p><p>TWO</p>"
let idx

// some code not shown here parses the first paragraph
// and updates idx
idx = 10

// next parse step must continue from idx 
let myRegex = /<p>[^<]*<\/p>/
let subbed = myString.substring(idx)
let result = myRegex.exec(subbed)
console.log(result) // "<p>TWO</p>", not "<p>ONE</p>"

But myString.substring(idx) seems like a quite expensive operation.

Are there no regex operations like this: result = myRegex.execFromIndex(idx, myString);?

In general, I want to start regex matching from different indexes so I can exclude parts of the string and avoid matches that are already parsed. So one time it can be from myString[0] another time myString[51] and so on.

Is there a way to do this efficiently? I'm parsing hundreds of thousands of lines and want to do this in an as cheap way as possible.

• javascript
• regex
• parsing

• 1 typo myString.length – Pranav C Balan Commented Jan 18, 2017 at 15:45
• 1 if you really are concerned about efficiency, try not to use regex. – Faibbus Commented Jan 18, 2017 at 15:47
• 4 Construct the regex instance and then set its .lastIndex property. Read the documentation. – Pointy Commented Jan 18, 2017 at 15:49
• @Faibbus What do you suggest instead? I'm quite sure I won't be able to write a more efficient search than regex on my own. – mottosson Commented Jan 18, 2017 at 15:54
• This question might get better answers if you would provide the logic by which you decide you need to search from a certain index only. – trincot Commented Jan 18, 2017 at 15:57

2 Answers 2

Use Regexp.exec and lastIndex

1. Create a Regexp with the y or g flag
   • with the y flag, the match must start exactly at the specified start index
   • with the g flag, the match can occur anywhere after the specified index
2. Set its lastIndex property to the start index
3. Call exec

I've applied the above steps to your example code:

let myString = "<p>ONE</p><p>TWO</p>"
let idx

// some code not shown here parses the first paragraph
// and updates idx
idx = 10

// next parse step must continue from idx 
let myRegex = /<p>[^<]*<\/p>/y  //