R help - May 2004 - tapply & hist

I'm learning how to use tapply. 
Now I'm having a go at the following code in which dati contains almost 600 
lines, Pot - numeric - are the capacities of power plants and SGruppo - text 
- the corresponding six technologies ("CCC",
"CIC","TGC", "CSC","CPC",
"TE").
.....................................................

dati=sqlQuery(canale,"select Id,SGruppo,Classe, NGruppo,ProdNetta,Pot from 
SintesiQuery")
attach(dati)
# Grouping by technology
tapply(Pot,SGruppo,sum)
...................................
# Histograms by technology
par(mfrow=c(2,3)) 
tapply(Pot,SGruppo,hist)
detach(dati)

It all works great but  tapply(Pot,SGruppo,hist) produces 6 histograms with 
the titles and the xlab labels in a generic form, something like integer[1], 
integer[2], ....... while I'd like to have each graph indicating the 
mentioned technologies.
I've been trying issuing 
tech=c("CCC", "CIC","TGC",
"CSC","CPC", "TE")
tapply(Pot,SGruppo,hist, main=tech)

but R prints in each histogram the six values in the title without cycling 
among them.

How can I obtain what I want?

Ciao
Vittorio 


to no avail

As another respondent already mentioned, Lattice is probably the way to
go on this one but if you do want to use tapply try this:

names(Pot) <- SGruppo
dummy <- tapply(Pot,SGruppo,function(x)hist(x,main=names(x)[1],xlab=NULL))


Vittorio <v.demartino2 <at> virgilio.it> writes:

: 
: I'm learning how to use tapply. 
: Now I'm having a go at the following code in which dati contains almost
600
: lines, Pot - numeric - are the capacities of power plants and SGruppo - text 
: - the corresponding six technologies 
("CCC", "CIC","TGC",
"CSC","CPC", "TE").
: .....................................................
: 
: dati=sqlQuery(canale,"select Id,SGruppo,Classe, NGruppo,ProdNetta,Pot
from
: SintesiQuery")
: attach(dati)
: # Grouping by technology
: tapply(Pot,SGruppo,sum)
: ...................................
: # Histograms by technology
: par(mfrow=c(2,3)) 
: tapply(Pot,SGruppo,hist)
: detach(dati)
: 
: It all works great but  tapply(Pot,SGruppo,hist) produces 6 histograms with 
: the titles and the xlab labels in a generic form, something like integer[1], 
: integer[2], ....... while I'd like to have each graph indicating the 
: mentioned technologies.
: I've been trying issuing 
: tech=c("CCC", "CIC","TGC",
"CSC","CPC", "TE")
: tapply(Pot,SGruppo,hist, main=tech)
: 
: but R prints in each histogram the six values in the title without cycling 
: among them.
: 
: How can I obtain what I want?
: 
: Ciao
: Vittorio

> ...................................
> # Histograms by technology
> par(mfrow=c(2,3))
> tapply(Pot,SGruppo,hist)
> detach(dati)
>
> It all works great but  tapply(Pot,SGruppo,hist) produces 6 histograms
> with
> the titles and the xlab labels in a generic form, something like
> integer[1],
> integer[2], ....... while I'd like to have each graph indicating the
tapply takes atomic data (usually vectors).  You want to pass rows of a
data frame, so the Pot *and* SGruppo will be sent together; "by()" is
very
good for this.  It might be possible (even easy?) to use tapply, but I
just use "by" for these things.

Since dati is your data frame, try this (untested!):

by(dati,dati$SGruppo, function(x,...){
  hist(x$Pot,main=as.character(x$SGruppo[1])) } )

Or, use Lattice:

library(lattice)
histogram( ~ Pot | SGruppo, data=dati)

Cheers

Jason