Fast Levenshtein distance in R?

Question

Is there a package that contains Levenshtein distance counting function which is implemented as a C or Fortran code? I have many strings to compare and stringMatch from MiscPsycho is too slow for this.

score 21 · Answer 1 · answered Mar 28 '14 at 16:45

21

And stringdist in the stringdist package does it too, even faster than levenshteinDist under certain conditions (1)

answered Mar 28 '14 at 16:45

Ben

41,615
18
132
227

3

stringdist has sped up significantly since that blog you link to: it now uses multiple cores. – Feb 26 '16 at 17:02

score 17 · Accepted Answer · edited Oct 03 '16 at 19:29

17

levenshteinDist (from the RecordLinkage package) calls compiled C code. Give it a try.

edited Oct 03 '16 at 19:29

MichaelChirico

33,841
14
113
198

answered Jul 05 '10 at 21:00

gd047

29,749
18
107
146

2

Just noting the RecordLinkage package is apparently no longer maintained and has been pulled from CRAN. The `stringdist` package is the solution now. – Brian Stamper Feb 27 '20 at 17:42
Just noting the RecordLinkage package is *not* pulled from CRAN, it’s just available: https://cran.r-project.org/web/packages/RecordLinkage/ – MS Berends Aug 12 '22 at 19:41

score 6 · Answer 3 · edited Oct 03 '16 at 19:30

6

You could try stringDist from Biostrings as well

edited Oct 03 '16 at 19:30

MichaelChirico

33,841
14
113
198

answered Jul 06 '10 at 02:37

Aaron Statham

2,048
1
15
16

score 1 · Answer 4 · answered Mar 08 '20 at 09:24

1

You could also use levenshtein_distance() from the textTinyR package. I got 'calloc' memory errors with all other packages when it came to larger character vectors of around 30k characters. Only textTinyR worked for me!

answered Mar 08 '20 at 09:24

interrobang

83
1
7

Fast Levenshtein distance in R?

4 Answers4

Linked

Related