Skip to content
Back to formatted view

Raw Message

Message-ID: <f8e6ff050809261454t1f6ea29bycde43b6d27a03a44@mail.gmail.com>
Date: 2008-09-26T21:54:10Z
From: Hadley Wickham
Subject: Newbie: Ranking a data frame, grouped by 2 or more columns
In-Reply-To: <82ba77b80809261354q47c21a64rb6ee185e28e6339a@mail.gmail.com>

On Fri, Sep 26, 2008 at 3:54 PM, Matthew Pettis
<matthew.pettis at gmail.com> wrote:
> Hi,
>
> I'd like to rank obs in a data frame as subset by 2 or more columns...
> The example input would look like the following:
>
> ====+====+====+====+
> x   y   v
> --  --  --
> a   w   200
> a   w   100
> b   w   500
> b   w   200
> b   z   300
> b   z   400
> ====+====+====+====+
>
> And the data frame I want to create is below:
> ====+====+====+====+
> x   y   v   rank
> --  --  --  ----
> a   w   200    1
> a   w   100    2
> b   w   500    1
> b   w   200    2
> b   z   300    2
> b   z   400    1
> ====+====+====+====+
>
> Can someone help me with this?

This is easy to do with the (very soon to be released) plyr package:

library(plyr)
ddply(df, .(x, y), transform, rank = rank(v))

You can learn more about it at http://had.co.nz/plyr

Hadley

-- 
http://had.co.nz/