Average within R data frame

Question

I have a dataframe that looks like this:

  name  val type group
1 A3330 184 ave IX13
2 A3330 42  min IX13
3 A1473 195 ave IX01
4 A1473 93  min IX01
5 A3330 80  ave BG12
6 A3330 44  min BG12
7 A1473 15  ave IX3
8 A1473 91  min IX3

I want to average every value in val, across groups if the type is the same. So 1 and 5 are averaged despite the fact that they are different groups. The same goes with 2 and 6, 3 and 7, and 4 and 8. In the end, I hope to have a new dataframe that looks like this:

name   val type
A3330  132 ave
A333   43  min
A1473  105 ave
A1473  92  min

I am thinking that I should do something like this:

for (metric in type)
 for(names in name)
  mean(df[df$name==names && type==metric, df$type])
 }
}

But in my limited R experience, there is always a better way than to loop. Any suggestions?

score 2 · Accepted Answer · answered Oct 09 '13 at 17:28

2

Haven't tested it myself but using the data.table package, something like this should work for you -

library(data.table)
df <- data.table(df)
df2 <- df[,
   list(
      meanval = mean(val)
   ),
   by = c('name','type')
]

answered Oct 09 '13 at 17:28

TheComeOnMan

12,535
8
39
54

I am getting this error: Error in `[.data.frame`(df, , list(meanval = mean("val")), by = c("name", : unused argument (by = c("name", "type")) Does this error mean anythign to you? – user2076476 Oct 09 '13 at 18:37
You might have missed `df <- data.table(df)` – TheComeOnMan Oct 09 '13 at 18:47
Darn. I did it but then forgot to use the new df. THANKS! – user2076476 Oct 09 '13 at 19:02

Average within R data frame

1 Answers1