CODEWITHSUNDEEP
rdataframeconditional-statements
Louis
Louis

Reputation: 1

Merging two dataframes into a new (on a condition)

I am new to R, so please bear with me.

I have two dataframes: 

df1 <- data.frame(name = c("name 1", "name 2", "name 3", "name 4"),
              columnname = c("hello", "", "hello", ""))

df2 <- data.frame(name = c("name 1", "name 2", "name 3"),
              columnname = c(1, 2, 3))

Which looks like:

df1
#name    columnname
#name 1      hello
#name 2           
#name 3      hello
#name 4      
df2
#name     columnname
#name 1          1
#name 2          2
#name 3          3

My goal is to replace the value "hello" in df1 with the corresponding value in df2 (and NA otherwise), and create a new dataframe, df3. So far I have the following code: 

fun <- function(cat_df, ret_df, col_name) {
ret_df[, col_name] <- ifelse(cat_df[, col_name] == "hello", ret_df[, col_name],"NA") 
return(ret_df)
}

df3 <- fun(df1, df2, col_name = "columnname")

df3
#name   columnname
#name 1          1
#name 2         NA
#name 3          3
#name 4         NA

However, I have 350 columns and 3000 rows. So my question is, how can I expand the code to hold a dataframe of 350 columns and 3000 rows? Other types of code are very welcome!

Upvotes: 0

Views: 60

Answers (2)

Louis
Louis

Reputation: 1

So if my two data frames have dimensions of: 

dim(df1)
639 260

and dim(df2)
2273 260

Would the code then look like:

 set.seed(4)
nobs=2273
df1 <- data.frame(name=paste("name",1:nobs))
df1[,paste0("col",1:260)] <- sample(c("hello",""),260*nobs,T)

df2 <- data.frame(name=paste("name",1:nobs))
df2[,paste0("col",1:260)] <- 1:nobs

mycols <- colnames(df1)[-1]
names(mycols) <- mycols
df3 <- data.frame(name=df1$name)
df3[mycols]<- lapply(mycols,function(x){
ifelse(df1[,x]=="hello",df2[,x],NA)
})

df3

?

Upvotes: 0

Heroka
Heroka

Reputation: 13149

Because you wanted a solution for multiple columns, we first create some data with multiple columns (you could've done this yourself....)

set.seed(4)
nobs=5
df1 <- data.frame(name=paste("name",1:nobs))
df1[,paste0("col",1:5)] <- sample(c("hello",""),5*nobs,T)
# name  col1  col2  col3  col4  col5
# 1 name 1       hello       hello      
# 2 name 2 hello       hello            
# 3 name 3 hello       hello            
# 4 name 4 hello                   hello
# 5 name 5       hello hello            

df2 <- data.frame(name=paste("name",1:nobs))
df2[,paste0("col",1:5)] <- 1:nobs
# name col1 col2 col3 col4 col5
# 1 name 1    1    1    1    1    1
# 2 name 2    2    2    2    2    2
# 3 name 3    3    3    3    3    3
# 4 name 4    4    4    4    4    4
# 5 name 5    5    5    5    5    5

Then we create a named vector of columns

mycols <- colnames(df1)[-1]
names(mycols) <- mycols

And make our results

df3 <- data.frame(name=df1$name)
df3[mycols]<- lapply(mycols,function(x){
  ifelse(df1[,x]=="hello",df2[,x],NA)
})


    name col1 col2 col3 col4 col5
1 name 1   NA    1   NA    1   NA
2 name 2    2   NA    2   NA   NA
3 name 3    3   NA    3   NA   NA
4 name 4    4   NA   NA   NA    4
5 name 5   NA    5    5   NA   NA

Upvotes: 1

Related Questions