CODEWITHSUNDEEP
rdataframemergerbind
tataṣe
tataṣe

Reputation: 59

rbind with partial column match?

I have two dataframes of the following form

df1:

first second
1 5
2 6
3 7

df2:

sec fir
3 4
2 1
1 3

I'm trying to use rbind (or anything else that works haha) to merge the rows of these two dataframes by using a partial match on the column names (matching first with fir, and second with sec) to create a dataframe that looks like this

first second
1 5
2 6
3 7
4 3
1 2
3 1

Thank you all in advance for the help!!!!

Upvotes: 0

Views: 208

Answers (2)

akrun
akrun

Reputation: 887048

With rbind, it needs the columns to have the same name. So, the option is to set the names of the second data with the names of the first one and then rbind

rbind(df1, setNames(df2, names(df1)))

If we need to automate by checking the substring, a distance based approach can be used

library(phonics)
nm1 <- c(names(df1), names(df2))
ind <- soundex(nm1, maxCodeLen = 2)
nm2 <- ave(nm1, ind, FUN = function(x) x[1])
rbind(df1, setNames(df2[match(ind[1:2], ind[3:4])], names(df1)))

Upvotes: 2

Ronak Shah
Ronak Shah

Reputation: 388907

You can use grep to find the column name in df2 that is similar to one in df1.

cols <- sapply(names(df2), function(x) grep(x, names(df1)))
cols
#sec fir 
#  2   1

Using cols you can rearrange df2, make their names similar to df1 and then rbind.

rbind(df1, setNames(df2[cols], names(df1)))

#  first second
#1     1      5
#2     2      6
#3     3      7
#4     4      3
#5     1      2
#6     3      1

Upvotes: 1

Related Questions