Can you set variables as a column name?

Question

If I have a dataframe as follows, with a combination of text values and NA cells:

id	Col1	Col2	Col3	Col4	Col5	Col...
id1	NA	NA	sample	NA	weight	etc
id2	NA	size	NA	NA	NA	etc
id3	volume	size	sample	NA	NA	etc
id4	NA	NA	NA	qty	NA	etc
id5	NA	NA	sample	qty	weight	etc

Is it possible to rename the header with the most frequently used value in that column, as follows?

id	volume	size	sample	qty	weight
id1	NA	NA	sample	NA	weight
id2	NA	size	NA	NA	NA
id3	volume	size	sample	NA	NA
id4	NA	NA	NA	qty	NA
id5	NA	NA	sample	qty	weight

score 2 · Accepted Answer · answered Mar 03 '21 at 01:27

Try the code below

> cbind(df[1], setNames(df[-1], sapply(df[-1], function(x) unique(na.omit(x)))))
   id volume size sample  qty weight
1 id1   <NA> <NA> sample <NA> weight
2 id2   <NA> size   <NA> <NA>   <NA>
3 id3 volume size sample <NA>   <NA>
4 id4   <NA> <NA>   <NA>  qty   <NA>
5 id5   <NA> <NA> sample  qty weight

Data

> dput(df)
structure(list(id = c("id1", "id2", "id3", "id4", "id5"), Col1 = c(NA,
NA, "volume", NA, NA), Col2 = c(NA, "size", "size", NA, NA),
    Col3 = c("sample", NA, "sample", NA, "sample"), Col4 = c(NA,
    NA, NA, "qty", "qty"), Col5 = c("weight", NA, NA, NA, "weight"
    )), class = "data.frame", row.names = c(NA, -5L))

score 1 · Answer 2 · answered Mar 03 '21 at 01:59

You can get the most frequent value in each column using Mode function from here.

Mode <- function(x) {
  ux <- unique(na.omit(x))
  ux[which.max(tabulate(match(x, ux)))]
}

Apply it for each column and change the column name.

names(df)[-1] <- sapply(df[-1], Mode)
df

#   id volume size sample  qty weight
#1 id1   <NA> <NA> sample <NA> weight
#2 id2   <NA> size   <NA> <NA>   <NA>
#3 id3 volume size sample <NA>   <NA>
#4 id4   <NA> <NA>   <NA>  qty   <NA>
#5 id5   <NA> <NA> sample  qty weight

Thanks, I didn't realise that there was a mode function that could handle text- I thought that it would only work on integers. — Harry, Mar 03 '21 at 13:59

Can you set variables as a column name?

2 Answers2