get row with max values in all columns in R

Question

Possible Duplicate:
Extracting indices for data frame rows that have MAX value for named field

hello,

I have a data frame like this :

   A1 A3    d
1   a pr    5
2   a be    0
3   a cd    8
4   a dy    0
5   b pr    3
6   b be    4
7   b cd    9

etc...

I want to test each row, and get the unique rows based on A1 and have max value of d

the output should be like this

A1 A3 d
a  cd 8
b  cd 9

etc..

The data frame is bigger , but that's an example.

Can this be done with R? without looping and long stuff??

thanks

@Joris Meys : it's a duplicate , but i did not understand the whole method used , i was able to generate all ids of the row that have max value , i wasn't able to get a subset of the original data frame based on this id, how this can be done?? — weblover, May 19 '11 at 11:47
first of all, read http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example/ thoroughly. Second, if you have the ids and you don't know how to subset (see eg ?subset, ?Extract, ...) , it's very much time to start reading any of http://stats.stackexchange.com/questions/138/resources-for-learning-r — Joris Meys, May 19 '11 at 11:51
@Joris Meys : thanks for your reply. i know how to extract them , but i needed another way . — weblover, May 19 '11 at 12:00

score 2 · Accepted Answer · answered May 19 '11 at 11:39

2

The easiest way to do it is to sort the d column, and them remove duplicates in the A1 column:

df2 <- df[order(df$d,decreasing=T),]
df2[!duplicated(df2$A1),]

This does assume that there is a single unique maximum, you would lose some results if there were more than 1.

answered May 19 '11 at 11:39

James

65,548
14
155
193

thanks alot , it worked :) , but i think there is more than 1 maximum value in each column , but i only need 1. – weblover May 19 '11 at 11:59

score 1 · Answer 2 · answered May 19 '11 at 11:40

1

Probably

ddply(dfr, "A1", function(curdfr){curdfr[which.max(curdfr$d),]})

answered May 19 '11 at 11:40

Nick Sabbe

11,684
1
43
57

1

Note: I haven't tested this as there wasn't proper example data provided (as in: easy to test with, without having to resort to textconnection and the like; read: `dput`.) – Nick Sabbe May 19 '11 at 11:41
+1 for comments that take long to type than to create sample data! – Brandon Bertelsen May 19 '11 at 15:15

score 0 · Answer 3 · answered May 19 '11 at 20:15

DATA

mydf <- read.table(textConnection("
 Lp   A1 A3    d
 1   a pr    5
 2   a be    0
 3   a cd    8
 4   a dy    0
 5   b pr    3
 6   b be    4
 7   b cd    9"),header=T,row.names="Lp")

CODE

require(data.table)
mydf <- data.table(mydf)
mydf[,.SD[which.max(d)],by=A1]

get row with max values in all columns in R

3 Answers3