[R] Doing operations by grouping variable

[Bill.Venables at csiro.au]

You left out the subscript.  Why not just do

d <- within(data.frame(group = rep(1:5, each = 5), variable = rnorm(25)), 
	scaled <- variable/tapply(variable, group, max)[group])

and be done with it?

(Warning: if you replace the second '<-' above by '=', it will not work.  
It is NOT true that you can always replace '<-' by '=' for assignment. Why?)

Bill Venables.

-----Original Message-----
From: Seth W Bigelow [mailto:sbigelow at fs.fed.us] 
Sent: Wednesday, 22 September 2010 1:43 AM
To: Venables, Bill (CMIS, Cleveland)
Cc: michael.bedward at gmail.com; r-help at r-project.org
Subject: RE: [R] Doing operations by grouping variable


Thanks, Bill and Michael, you have answered the question I asked, but not the one I wished to ask I want to obtain the maximum in each group of variables, so I could scale each variable by the maximum for its group. If I use tapply, as in the example below, there's a mismatch in dimensions of the output of tapply [5] and the data frame with the variables[25]. 


group <- rep(1:5, each=5)             # define grouping variable                                 

variable <- rnorm(25)                 # generate data 

d <- data.frame(group,variable)       # bundle together in a data frame 

d$scaled <- d$variable/(with(d,tapply(variable,group,max)))    # crash and burn 





Dr. Seth  W. Bigelow
Biologist, USDA-FS Pacific Southwest Research Station
1731 Research Park Drive, Davis California




<Bill.Venables at csiro.au> 

09/20/2010 06:24 PM 

	
To
	<michael.bedward at gmail.com>, <sbigelow at fs.fed.us>, <r-help at r-project.org> 
cc
	
Subject
	RE: [R] Doing operations by grouping variable

	




That's if the variables are visible.  If they are only in the data frame it's not much more difficult

d <- data.frame(group = rep(1:5, each=5), 
               variable = rnorm(25))
with(d, tapply(variable, group, max))


(Tip: avoid using attach().)

Bill Venables.  

-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] On Behalf Of Michael Bedward
Sent: Tuesday, 21 September 2010 11:15 AM
To: Seth W Bigelow; Rhelp
Subject: Re: [R] Doing operations by grouping variable

Not sure why you think tapply is "awkward". Your example would be...

group <- rep(1:5, each=5)
variable <- rnorm(25)
tapply(variable, group, max)

...which looks quite elegant to me :)

Meanwhile, the reason your expression doesn't work is that you are
asking mistakenly for elements 1:5 repeatedly from the variable col.
If you just type d$variable[ d$group ] and compare the values to your
variable vector this should be clear.

Michael

On 21 September 2010 10:59, Seth W Bigelow <sbigelow at fs.fed.us> wrote:
> I'm writing an expression that requires searching a vector according to
> group. As an example, I want to find the maximum value in each of 5
> groups.
>
>
> group=rep(1:5, each=5)                          # create grouping variable
>
> variable=rnorm(25)                              # generate data
>
> d <- data.frame(group,variable)                         # make data frame
>
> max(d$variable[d$group])                        # try expression that
> doesn't work
>
> I'm expecting a vector containing the maximum variable value, per group.
> What am I doing wrong? I know I can use aggregate, tapply, etc. but that
> seems awkward and bulky, is there a simpler way?
>
>
> Dr. Seth  W. Bigelow
> Biologist, USDA-FS Pacific Southwest Research Station
> 1731 Research Park Drive, Davis California
>
>        [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide <https://stat.ethz.ch/mailman/listinfo/r-help> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

______________________________________________
R-help at r-project.org mailing list
<http://www.r-project.org/posting-guide.html> https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide <https://stat.ethz.ch/mailman/listinfo/r-help> http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
<http://www.r-project.org/posting-guide.html>