Summarizing factor data in table?

-----Original Message-----
From: Tony Plate [mailto:tplate at acm.org]
Sent: Tuesday, April 26, 2005 2:00 PM
To: Andy Bunn
Cc: R-Help
Subject: Re: [R] Summarizing factor data in table?


Do you want to count the number of non-NA divisions and organizations in
the data for each year (where duplicates are counted as many times as
they appear)?

 > tapply(!is.na(foo$div), foo$yr, sum)

1998 1999 2000
    0    4    2

 > tapply(!is.na(foo$org), foo$yr, sum)

1998 1999 2000
    4    4    2

 >

Or perhaps the number of unique non-NA divisions and organizations in
the data for each year?

 > tapply(foo$div, foo$yr, function(x) length(na.omit(unique(x))))

1998 1999 2000
    0    4    2

 > tapply(foo$org, foo$yr, function(x) length(na.omit(unique(x))))

1998 1999 2000
    4    4    2

 >

(I don't understand where the "3" in your desired output comes from
though, which maybe indicates I completely misunderstand your request.)

Andy Bunn wrote:

I have a very simple query with regard to summarizing the

number of factors

present in a certain snippet of a data frame.
Given the following data frame:

	foo <- data.frame(yr = c(rep(1998,4), rep(1999,4),

rep(2000,2)), div =

factor(c(rep(NA,4),"A","B","C","D","A","C")),
      	            org = factor(c(1:4,1:4,1,2)))

I want to get two new variables. Object ndiv would give the number of
divisions by year:
     1998 0
     1999 3
     2000 2
Object norgs would give the number of organizations
     1998 4
     1999 4
     2000 2
I figure xtabs should be able to do it, but I'm stuck without a

for loop.

Any suggestions? -Andy