R help - May 2005 - Contingency tables from data.frames

Dear list,

I'm trying to do a set of generic functions do make contingency tables from 
data.frames. It is just running "nice" (I'm learning R), but I
think it can be
better.

I would like to filter the data.frame, i.e, eliminate all not numeric variables.
And I don't know how to make it: please, help me.

Below one of the my functions ('er' is a mention to EasieR, because
I'm trying
to do a package for myself and the my students):

#2. Tables from data.frames
#2.1---er.table.df.br (User define breaks and right)------------
er.table.df.br <- function(df,
                            breaks = c('Sturges', 'Scott',
'FD'),
                            right = FALSE) {

   if (is.data.frame(df) != 'TRUE')
     stop('need "data.frame" data')

   dim_df <- dim(df)

   tmpList <- list()

   for (i in 1:dim_df[2]) {

     x <- as.matrix(df[ ,i])
     x <- na.omit(x)

     k <- switch(breaks[1],
                 'Sturges' = nclass.Sturges(x),
                 'Scott'   = nclass.scott(x),
                 'FD'      = nclass.FD(x),
                 stop("'breaks' must be 'Sturges',
'Scott' or 'FD'"))

     tmp      <- range(x)
     classIni <- tmp[1] - tmp[2]/100
     classEnd <- tmp[2] + tmp[2]/100
     R        <- classEnd-classIni
     h        <- R/k

     # Absolut frequency
     f <- table(cut(x, br = seq(classIni, classEnd, h), right = right))

     # Relative frequency
     fr <- f/length(x)

     # Relative frequency, %
     frP <- 100*(f/length(x))

     # Cumulative frequency
     fac <- cumsum(f)

     # Cumulative frequency, %
     facP <- 100*(cumsum(f/length(x)))

     fi   <- round(f, 2)
     fr   <- round(as.numeric(fr), 2)
     frP  <- round(as.numeric(frP), 2)
     fac  <- round(as.numeric(fac), 2)
     facP <- round(as.numeric(facP),2)

     # Table
     res <- data.frame(fi, fr, frP, fac, facP)
     names(res) <- c('Class limits', 'fi', 'fr',
'fr(%)', 'fac', 'fac(%)')
     tmpList <- c(tmpList, list(res))
   }
   names(tmpList) <- names(df)
   return(tmpList)
}

To try the function:

#a) runing nice
y1=rnorm(100, 10, 1)
y2=rnorm(100, 58, 4)
y3=rnorm(100, 500, 10)
mydf=data.frame(y1, y2, y3)
#tbdf=er.table.df.br (mydf, breaks = 'Sturges', right=F)
#tbdf=er.table.df.br (mydf, breaks = 'Scott', right=F)
tbdf=er.table.df.br (mydf, breaks = 'FD', right=F)
print(tbdf)


#b) One of the problems
y1=rnorm(100, 10, 1)
y2=rnorm(100, 58, 4)
y3=rnorm(100, 500, 10)
y4=rep(letters[1:10], 10)
mydf=data.frame(y1, y2, y3, y4)
tbdf=er.table.df.br (mydf, breaks = 'Scott', right=F)
print(tbdf)

Could anyone give me a hint how to work around this?

PS: Excuse my bad English ;-)
-- 
Jose Claudio Faria
Brasil/Bahia/UESC/DCET
Estatistica Experimental/Prof. Adjunto
mails:
  joseclaudio.faria at terra.com.br
  jc_faria at uesc.br
  jc_faria at uol.com.br

On 5/24/05, Jose Claudio Faria <joseclaudio.faria at terra.com.br>
wrote:
> Dear list,
> 
> I'm trying to do a set of generic functions do make contingency tables
from
> data.frames. It is just running "nice" (I'm learning R), but
I think it can be
> better.
> 
> I would like to filter the data.frame, i.e, eliminate all not numeric
variables.
> And I don't know how to make it: please, help me.
> 
> Below one of the my functions ('er' is a mention to EasieR, because
I'm trying
> to do a package for myself and the my students):
> 
> #2. Tables from data.frames
> #2.1---er.table.df.br (User define breaks and right)------------
> er.table.df.br <- function(df,
>                            breaks = c('Sturges', 'Scott',
'FD'),
>                            right = FALSE) {
> 
>   if (is.data.frame(df) != 'TRUE')
>     stop('need "data.frame" data')
> 
>   dim_df <- dim(df)
> 
>   tmpList <- list()
> 
>   for (i in 1:dim_df[2]) {
> 
>     x <- as.matrix(df[ ,i])
>     x <- na.omit(x)
> 
>     k <- switch(breaks[1],
>                 'Sturges' = nclass.Sturges(x),
>                 'Scott'   = nclass.scott(x),
>                 'FD'      = nclass.FD(x),
>                 stop("'breaks' must be 'Sturges',
'Scott' or 'FD'"))
> 
>     tmp      <- range(x)
>     classIni <- tmp[1] - tmp[2]/100
>     classEnd <- tmp[2] + tmp[2]/100
>     R        <- classEnd-classIni
>     h        <- R/k
> 
>     # Absolut frequency
>     f <- table(cut(x, br = seq(classIni, classEnd, h), right = right))
> 
>     # Relative frequency
>     fr <- f/length(x)
> 
>     # Relative frequency, %
>     frP <- 100*(f/length(x))
> 
>     # Cumulative frequency
>     fac <- cumsum(f)
> 
>     # Cumulative frequency, %
>     facP <- 100*(cumsum(f/length(x)))
> 
>     fi   <- round(f, 2)
>     fr   <- round(as.numeric(fr), 2)
>     frP  <- round(as.numeric(frP), 2)
>     fac  <- round(as.numeric(fac), 2)
>     facP <- round(as.numeric(facP),2)
> 
>     # Table
>     res <- data.frame(fi, fr, frP, fac, facP)
>     names(res) <- c('Class limits', 'fi', 'fr',
'fr(%)', 'fac', 'fac(%)')
>     tmpList <- c(tmpList, list(res))
>   }
>   names(tmpList) <- names(df)
>   return(tmpList)
> }
> 
> To try the function:
> 
> #a) runing nice
> y1=rnorm(100, 10, 1)
> y2=rnorm(100, 58, 4)
> y3=rnorm(100, 500, 10)
> mydf=data.frame(y1, y2, y3)
> #tbdf=er.table.df.br (mydf, breaks = 'Sturges', right=F)
> #tbdf=er.table.df.br (mydf, breaks = 'Scott', right=F)
> tbdf=er.table.df.br (mydf, breaks = 'FD', right=F)
> print(tbdf)
> 
> 
> #b) One of the problems
> y1=rnorm(100, 10, 1)
> y2=rnorm(100, 58, 4)
> y3=rnorm(100, 500, 10)
> y4=rep(letters[1:10], 10)
> mydf=data.frame(y1, y2, y3, y4)
> tbdf=er.table.df.br (mydf, breaks = 'Scott', right=F)
> print(tbdf)
> 
Try this:

sapply(my.data.frame, is.numeric)

Also you might want to look up:

?match.arg
?stopifnot
?ncol
?sapply
?lapply