6

Is it possible to use spaces in ddply?

I'm using data from a spreadsheet with a lot of spaces in column names and i would like to keep those names because later on I want to export this data with the same column names as the original. There are 200+ columns and using make.names will of course give me proper names but then I lose the original column names.

However ddply doesn't seem to like spaces? Is there a workaround?

lev=gl(2, 3, labels=c("low", "high"))
df=data.frame(factor=lev, "fac tor"=lev, response=1:6, check.names = FALSE)

> ddply(df, c("factor"), summarize, r.avg=mean(response))
factor r.avg
1    low     2
2   high     5

> ddply(df, c("fac tor"), summarize, r.avg=mean(response))
Error in parse(text = x) : <text>:1:5: unexpected symbol
: fac tor
Johan
  • 810
  • 6
  • 12
  • 1
    I think you meant `check.names = FALSE` up above. `TRUE` forces coercion to `fac.tor` – Chase Jul 04 '11 at 08:42

2 Answers2

9

Wrapping the column names in single back ticks (`) seems to do the trick.

ddply(df, "`fac tor`", summarize, r.avg=mean(response))

You can also use column indices which may or may not be appealing depending on how big your data.frame is and your knowledge of the locations of each column beforehand.

ddply(df, 2, summarize, r.avg=mean(response))
Chase
  • 67,710
  • 18
  • 144
  • 161
  • +1 Yes, using backticks is the way to go. FYI, plyr provides the `.()` function that allows you to specify column names without having to quote them. So I would prefer writing `ddply(df, .(\`fac tor\`), summarize, r.avg=mean(response))` – Andrie Jul 04 '11 at 09:30
1

I would just use a regular expression to convert the spaces to some nonsense character, then convert back at the end:

lev=gl(2, 3, labels=c("low", "high"))
df=data.frame(factor=lev, "fac tor"=lev, response=1:6, check.names = FALSE)
colnames(df) <- gsub(" ","~",colnames(df))
Ari B. Friedman
  • 71,271
  • 35
  • 175
  • 235