R issues with merge/rbind/concatenate two data frames - r

Question

I am a beginner with R so i apologise in advance if the question was asked elsewhere. Here is my issue:
I have two data frames, df1 and df2, with different number of rows and columns. The two frames have only one variable (column) in common called "customer_no". I want the merged frame to match records based on "customer_no" and by rows in df2 only.Both data.frames have multiple rows for each customer_no.
I tried the following:
merged.df <- (df1, df2, by="customer_no",all.y=TRUE)
The problem is that this assigns values of df1 to df2 where instead it should be empty. My questions are:
1) How can I tell the command to leave the unmatched columns empty?
2) How can I see from the merged file which row came from which df? I guess if I resolve the above question this should be easy to see by the empty columns.
I am missing something in my command but don't know what. If the question has been answered somewhere else, would you be still kind enough to rephrase it in English here for an R beginner?
Thanks!
Data example:
df1:
customer_no country year
10 UK 2001
10 UK 2002
10 UK 2003
20 US 2007
30 AU 2006
df2:
customer_no income
10 700
10 800
10 900
30 1000
Merged file should look like this:
merged.df:
customer_no income country year
10 UK 2001
10 UK 2002
10 UK 2003
10 700
10 800
10 900
30 AU 2006
30 1000
So:
It puts the columns all together, it adds the values of df2 right after the last one of df1 based on same customer_no and matches only customer_no from df2 (merged.df does not have customer_no 20). Also, it leaves empty all the other cells.
In STATA I use append but not sure in R...perhaps join?
THANKS!!

Answer 1

Try:
df1$id <- paste(df1$customer_no, 1, sep="_")
df2$id <- paste(df2$customer_no, 2, sep="_")
res <- merge(df1, df2, by=c('id', 'customer_no'),all=TRUE)[,-1]
res1 <- res[res$customer_no %in% df2$customer_no,]
res1
# customer_no country year income
#1 10 UK 2001 NA
#2 10 UK 2002 NA
#3 10 UK 2003 NA
#4 10 <NA> NA 700
#5 10 <NA> NA 800
#6 10 <NA> NA 900
#8 30 AU 2006 NA
#9 30 <NA> NA 1000
If you want to change NA to '',
res1[is.na(res1)] <- '' #But, I would leave it as `NA` as there are `numeric` columns.
Or, use rbindlist from data.table (Using the original datasets)
library(data.table)
indx <- df1$customer_no %in% df2$customer_no
rbindlist(list(df1[indx,], df2),fill=TRUE)[order(customer_no)]
# customer_no country year income
#1: 10 UK 2001 NA
#2: 10 UK 2002 NA
#3: 10 UK 2003 NA
#4: 10 NA NA 700
#5: 10 NA NA 800
#6: 10 NA NA 900
#7: 30 AU 2006 NA
#8: 30 NA NA 1000

Answer 2

You could also use the smartbind function from the gtools package.
require(gtools)
res <- smartbind(df1[df1$customer_no %in% df2$customer_no, ], df2)
res[order(res$customer_no), ]
# customer_no country year income
# 1:1 10 UK 2001 NA
# 1:2 10 UK 2002 NA
# 1:3 10 UK 2003 NA
# 2:1 10 <NA> NA 700
# 2:2 10 <NA> NA 800
# 2:3 10 <NA> NA 900
# 1:4 30 AU 2006 NA
# 2:4 30 <NA> NA 1000

Answer 3

Try:
df1$income = df2$country = df2$year = NA
rbind(df1, df2)
customer_no country year income
1 10 UK 2001 NA
2 10 UK 2002 NA
3 10 UK 2003 NA
4 20 US 2007 NA
5 30 AU 2006 NA
6 10 <NA> NA 700
7 10 <NA> NA 800
8 10 <NA> NA 900
9 30 <NA> NA 1000