R help - May 2010 - Any way to apply TWO functions with tapply()?

I need to compute the mean and the standard deviation of a data set and would
like to have the results in one table/data frame. I call tapply() two times
and do then merge the resulting tables to have them all in one table. Is
there any way to tell tapply() to use the functions mean and sd within one
function call? Something like tapply(data$response, list(data$targets,
data$conditions), c(mean, sd)).

Thanks in advance.
-- 
View this message in context:
http://r.789695.n4.nabble.com/Any-way-to-apply-TWO-functions-with-tapply-tp2133924p2133924.html
Sent from the R help mailing list archive at Nabble.com.

Look at the reshape package - it allows that when using "cast"



----------------Contact
Details:-------------------------------------------------------
Contact me: Tal.Galili@gmail.com |  972-52-7275845
Read me: www.talgalili.com (Hebrew) | www.biostatistics.co.il (Hebrew) |
www.r-statistics.com (English)
----------------------------------------------------------------------------------------------




On Fri, May 7, 2010 at 12:39 PM, Phil Wieland <phwiel@gmx.de> wrote:
>
> I need to compute the mean and the standard deviation of a data set and
> would
> like to have the results in one table/data frame. I call tapply() two times
> and do then merge the resulting tables to have them all in one table. Is
> there any way to tell tapply() to use the functions mean and sd within one
> function call? Something like tapply(data$response, list(data$targets,
> data$conditions), c(mean, sd)).
>
> Thanks in advance.
> --
> View this message in context:
>
http://r.789695.n4.nabble.com/Any-way-to-apply-TWO-functions-with-tapply-tp2133924p2133924.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-help@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
	[[alternative HTML version deleted]]

Hi,

What you can do is define your own function which takes a vector of
values, computes the statistics you want and then returns a string
which displays the output the way you want it. Then use this function
in your tapply call.

like (untested)

mySummary <- function(x) {

  paste(mean(x),sd(x),sep=",")

}

tapply(data$response, list(data$targets, data$conditions) ,mySummary)

Of course, if you need a different output format, then you'll have to
adapt the paste call.

/Fredrik

On Fri, May 7, 2010 at 11:39 AM, Phil Wieland <phwiel at gmx.de>
wrote:
>
> I need to compute the mean and the standard deviation of a data set and
would
> like to have the results in one table/data frame. I call tapply() two times
> and do then merge the resulting tables to have them all in one table. Is
> there any way to tell tapply() to use the functions mean and sd within one
> function call? Something like tapply(data$response, list(data$targets,
> data$conditions), c(mean, sd)).
>
> Thanks in advance.
> --
> View this message in context:
http://r.789695.n4.nabble.com/Any-way-to-apply-TWO-functions-with-tapply-tp2133924p2133924.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>


-- 
"Life is like a trumpet - if you don't put anything into it, you
don't
get anything out of it."

tapply does handle functions with vector outputs, e.g. using the built
in CO2 data set the following data frame is returned:
> f <- function(x) data.frame(mean = mean(x), sd = sd(x))
> do.call(rbind, tapply(CO2$uptake, CO2$Type, f))
            mean   sd
Quebec      33.5 9.67
Mississippi 20.9 7.82

Note that if you replace data.frame with c in f then you get a matrix
out instead of a data.frame.

There is also somewhat similar functionality in summaryBy (doBy
package), summary.formula (Hmisc package), ddply (plyr package), remix
(remix package), melt and cast (reshape package) and sqldf (sqldf
package).

Of these summaryBy in the doBy package is particularly easy to specify
and produces a data frame:

library(doBy)
summaryBy(uptake ~ Type, data = CO2, FUN = c(mean, sd))

remix and summary.formula in Hmisc have particularly attractive output
but do not produce data frames.  Hmisc even has a plot method.  The
specification to remix is also simple here and, in fact, is identical
to the summaryBy line above except it uses lower case fun.  sqldf uses
SQL for the specification which may be an advantage if you know SQL
better than R.

On Fri, May 7, 2010 at 5:39 AM, Phil Wieland <phwiel at gmx.de>
wrote:
>
> I need to compute the mean and the standard deviation of a data set and
would
> like to have the results in one table/data frame. I call tapply() two times
> and do then merge the resulting tables to have them all in one table. Is
> there any way to tell tapply() to use the functions mean and sd within one
> function call? Something like tapply(data$response, list(data$targets,
> data$conditions), c(mean, sd)).
>
> Thanks in advance.
> --
> View this message in context:
http://r.789695.n4.nabble.com/Any-way-to-apply-TWO-functions-with-tapply-tp2133924p2133924.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

As pointed out to me offline, data.table should be added to the list
of relevant packages as well.  Its primary advantage is for large data
sets as it is very fast.  Its interface does take some getting used to
but its most recent version on CRAN does have several vignettes which
should ease learning.

On Fri, May 7, 2010 at 8:50 AM, Gabor Grothendieck
<ggrothendieck at gmail.com> wrote:
> tapply does handle functions with vector outputs, e.g. using the built
> in CO2 data set the following data frame is returned:
>
>> f <- function(x) data.frame(mean = mean(x), sd = sd(x))
>> do.call(rbind, tapply(CO2$uptake, CO2$Type, f))
> ? ? ? ? ? ?mean ? sd
> Quebec ? ? ?33.5 9.67
> Mississippi 20.9 7.82
>
> Note that if you replace data.frame with c in f then you get a matrix
> out instead of a data.frame.
>
> There is also somewhat similar functionality in summaryBy (doBy
> package), summary.formula (Hmisc package), ddply (plyr package), remix
> (remix package), melt and cast (reshape package) and sqldf (sqldf
> package).
>
> Of these summaryBy in the doBy package is particularly easy to specify
> and produces a data frame:
>
> library(doBy)
> summaryBy(uptake ~ Type, data = CO2, FUN = c(mean, sd))
>
> remix and summary.formula in Hmisc have particularly attractive output
> but do not produce data frames. ?Hmisc even has a plot method. ?The
> specification to remix is also simple here and, in fact, is identical
> to the summaryBy line above except it uses lower case fun. ?sqldf uses
> SQL for the specification which may be an advantage if you know SQL
> better than R.
>
> On Fri, May 7, 2010 at 5:39 AM, Phil Wieland <phwiel at gmx.de>
wrote:
>>
>> I need to compute the mean and the standard deviation of a data set and
would
>> like to have the results in one table/data frame. I call tapply() two
times
>> and do then merge the resulting tables to have them all in one table.
Is
>> there any way to tell tapply() to use the functions mean and sd within
one
>> function call? Something like tapply(data$response, list(data$targets,
>> data$conditions), c(mean, sd)).
>>
>> Thanks in advance.
>> --
>> View this message in context:
http://r.789695.n4.nabble.com/Any-way-to-apply-TWO-functions-with-tapply-tp2133924p2133924.html
>> Sent from the R help mailing list archive at Nabble.com.
>>
>> ______________________________________________
>> R-help at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
>>
>