- Issue created by @amitlokulwar
- 🇺🇸United States bkelly
Hi @amitlokulwar -
I had considered this but originally decided against it because of all the random numbers that would be generated for the the index.
Since ngrams are created by strings of letters bordering non letters. If this was extended to numbers then something as simple as a date would generate several more or less useless strings. Any set of numbers with a non alphanumeric delimiter might have a negative impact on results.
Example:
1 234-567-8901
Ngrams generated when limited to 3 or more characters.
234
234 567
234 567 8901
567
567 234
567 234
567 8901
8901 567
8901 567 234
8910Ngrams generated when limited to 4 or more characters.
8901The code is simple enough that it could be an option for those that want it.
I'll think about it.
Thank you for your time.
-- Bill Kelly