Odtnhj

I have a function that returns a data frame with more than 50 columns (one row).
This function is called inside a loop (lapply) so I get a list of dataframes one row each.

I initialize an empty dataframe at the very beginning in the function to be populated with the actual data returned by the process (short version here for testing)

 df <- read.csv(text = "conn,pc,server,profile")

R> df
[1] conn pc server profile car id amount 
<0 rows> (or 0-length row.names)

I would like to assign values to the row taking into account that the process sometimes does not return values for all columns.

Similar SO questions tend to rely on initializing data frame this way:

df <- data.frame("conn"= NA, "pc"= NA, "server"= NA, "profile" = NA)

but I don't want a row of NAs in case there's no value returned.

So in my function I tried first:

returned.conn.value <- 3 # assigned by hand for the sake of simplicity

df$conn <- returned.conn.value
 Error in `$<-.data.frame`(`*tmp*`, conn, value = 3) : 
 replacement has 1 row, data has 0

So I discovered there's no way of adding value to a column inside an empty dataframe, but.....

df[1,1] <- 3

This, to my surprise, worked. But, I consider very ugly to assign values to columns by index, so I tried this:

grep("^conn$", colnames(df))
[1] 1

Uglier indeed, as it's forcing me to write the very same line for every column (instead of counting by hand, yes)

So, I'm afraid I took a dead road leading to nowhere.

is there any way of initialize an empty dataframe as it will allow me to do simple assignments as the one below with no issue?

df$conn <- 3

Any hint will be much appreciated.