Let's look at the dedup function invocation once again:

> rec.pairs <- compare.dedup(RLdata500
+                        ,blockfld = list(1, 5:7)
+                        ,strcmp =   c(2,3,4)
+                        ,strcmpfun = levenshteinSim).

The strcmp and strcmpfun parameters dictate on which fields we need to do string comparison and what kind of string comparison we need to apply. We pass a vector indicating the column IDs to strcmp. We need to do string comparisons in columns 2, 3, and 4. We want to use the Levenshtein distance to find the similarity between two strings.

Let's look at the records 1 and 174; we see the first name matching, but no match with the rest of the fields. The Levenshtein distance of 0.142857 also states how far the last names are from each other:

> RLdata500[1,]
  fname_c1 fname_c2 lname_c1 lname_c2   by bm bd
1  CARSTEN     <NA>    MEIER     <NA> 1949  7 22
> RLdata500[174,]
    fname_c1 fname_c2 lname_c1 lname_c2   by bm bd
174  CARSTEN     <NA>  SCHMITT     <NA> 2001  6 27

We need to state the exact columns where string comparisons should be applied. Not specifying this may lead to unexpected results:

> rec.pairs.matches <- compare.dedup(RLdata500
+                                    ,blockfld = list(1, 5:7)
+                                    ,strcmp =   TRUE
+                                    ,strcmpfun = levenshteinSim)
> head(rec.pairs.matches$pairs)
  id1 id2 fname_c1 fname_c2  lname_c1 lname_c2   by  bm  bd is_match
1   1 174        1       NA 0.1428571       NA 0.00 0.5 0.5       NA
2   1 204        1       NA 0.0000000       NA 0.50 0.5 0.0       NA
3   2   7        1       NA 0.3750000       NA 0.75 0.5 0.0       NA
4   2  43        1       NA 0.8333333       NA 1.00 1.0 1.0       NA
5   2 169        1       NA 0.0000000       NA 0.50 0.0 0.5       NA
6   4  19        1       NA 0.1428571       NA 0.50 0.5 0.0       NA

You can see that the function has also calculated the string comparisons for the date of birth fields!