apply function on each data frames names stored in character vector

user_4246716 0
Viewed 50

I have data frames in my environment, and I have stored the names of the same in a character vector.

ldf <- names(which(unlist(eapply(.GlobalEnv,is.data.frame))))

when I print ldf, it has a structure similar to this:

>ldf
[1] "tdf1"           "tdf2"
[3] "tdf3"

I want to remove rows that has many na's from each of these data frames. I have created a function as follows:

remna <- function(df) {
  df$countofna <- apply(df[,c(1:length(df))], 1,function(x) sum(is.na(x)))
  df <- df[df$countofna <=3, ]
  df$countofna <- NULL
  df <- df[1:nrow(df)-1, ]
}

How do i apply the remna function on each of the dataframes with names stored in ldf. Without using ldf, I tried using a for loop: 

for (i in names(which(unlist(eapply(.GlobalEnv,is.data.frame))))) {
  remna(i)
}

however this fails at executing remna with the following error:

Error in df[, c(1:length(df))] : incorrect number of dimensions

The function remna works perfectly on individual data frames.

apply r
1 Answers

One option is to load the datasets in a list with mget and use rowSums to create a logical vector for keeping/removing the rows from each of the dataset

out <- lapply(mget(ldf), function(dat) dat[rowSums(is.na(dat)) <= 3,, drop = FALSE])

If the original objects needs to be updated, then use list2env (not recommended though)

list2env(out, .GlobalEnv)
user_37322710
Related