duplicates in agrep function

Question

I have the following code:

x <- data.frame("SN" = 1:2, "Name" = c("aaa","bbb"))

y <- data.frame("SN" = 1:2,  "Name" = c("aa1","aa2"))

x$partials<- as.character(sapply(x$Name, agrep, y$Name,max.distance = 1,value=T))

x

The output is the following:

    > x
  SN Name        partials
1  1  aaa c("aa1", "aa2")
2  2  bbb    character(0)

However I am expecting the following output:

Any ideas?

Please check https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example and fix your question — s_baldur, Feb 18 '19 at 10:27

jay.sf · Accepted Answer · 2019-02-18T11:00:42.217

You are probably looking for this.

First, the sapply() puts out nothing if character(0). To prevent this, you could say it's NA or text "character(0)" if you really want this.

z <- setNames(sapply(x$Name, function(a) {
  ag <- agrep(a, y$Name, max.distance=1, value=TRUE)
  if (identical(ag, character(0))) NA  # set to "character(0)" at will
  else ag
  }), x$Name)

Then, transform the list you get into a matrix.

z <- do.call(rbind, z)

We need to melt() it to get the right format. A good way is with data.table.

library(data.table)
z <- setNames(melt(z)[-2], c("Name", "partials"))

Now, we just merge x with the new data to get the result, ensuring unique rows of z.

res <- merge(x, unique(z))[c(2, 1, 3)]

> res
  SN Name partials
1  1  aaa      aa1
2  1  aaa      aa2
3  2  bbb     <NA>

duplicates in agrep function

1 Answers1