cross-tabulations on multiple data.frames and columns

Question

I'm calculating summary statistics for numerous data frames across multiple slices vs a single response variable. I currently do this by passing a list of DFs to a function. But my function has to specify the columns (ie-slices) individually. This speeds up my process dramatically; but, I think there has to be an even more efficient way to do this via an apply() family function. I'm hoping someone here can help me out.

Here's my code:

table1 <- function(x) {
  dl2 <- list()
  for (i in 1:length(x)) {
    z <- x[[i]]
    t.sliceA     <- addmargins(table(list(z$sliceA, z$Growing)))
    t.sliceB     <- addmargins(table(list(z$sliceB, z$Growing)))
    t.sliceC     <- addmargins(table(list(z$sliceC, z$Growing)))
    t.sliceD     <- addmargins(table(list(z$sliceD, z$Growing)))
    ...
    t.sliceAA    <- addmargins(table(list(z$sliceAA, z$Growing)))
    table.list <- list(t.sliceA, t.sliceB, t.sliceC, ... , t.sliceAA)
    names(table.list) <- c("t.sliceA", "t.sliceB", ... , "t.sliceAA")
    dl2[[i]] <- table.list
  }
  assign("dl",dl2, envir=.GlobalEnv)
}
# run the function
dl <- c(DF1, DF2, ..., DF.n)
table1(dl)

I assume there must be a more efficient way to do this via lapply() where I only have to specify the columns needed. Something where I would replace the lines

t.sliceA <- [blah]
...
t.sliceAA <- [blah]

with something like:

apply(z[,c(1:4,10:12,15)],2, function(x) addmargins(table(list(x,z$Growing))))

Any help that you can provide would be very helpful. Thanks!

Update: Reproducible example @Chase My apologies if the this was done poorly. It's my first time using github.

https://gist.github.com/3719220

and here's the code:

# load the example datasets
a.small <- dget("df1.txt")
l.small <- dget(df2.txt)

# working function that I'd like to simplify
table1 <- function(x) {
  dl2 <- list()
  for (i in 1:length(x)) {
    z <- x[[i]]
    t.tenure     <- addmargins(table(list(z$Tenure.Group, z$Growing)))
    t.optfile    <- addmargins(table(list(z$opt.file, z$Growing)))
    t.checking   <- addmargins(table(list(z$checking, z$Growing)))
    t.full      <- addmargins(table(list(z$add.full, z$Growing)))
    t.optdm      <- addmargins(table(list(z$opt.dm, z$Growing)))
    t.up         <- addmargins(table(list(z$add.up, z$Growing)))
    t.off        <- addmargins(table(list(z$offmode, z$Growing)))
    table.list <- list(t.tenure, t.optfile, t.checking, t.full, t.optdm, t.up, t.off)
    names(table.list) <- c("t.tenure", "t.optfile", "t.checking", "t.full", "t.optdm", "t.up", "t.off")
    dl2[[i]] <- table.list
  }
  assign("dl",dl2, envir=.GlobalEnv)
}
# create a DF list to send to the function
dl <- list(a.small, l.small)
table1(dl) # run the function

[reproducible](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) examples make the world go round and round, round and round... — Chase, Sep 13 '12 at 22:51

mnel · Accepted Answer · 2012-09-26T03:43:49.910

As far as I can see this will be easily done with a couple of lapply statements

If we define our function to create a table with margins as

tabulate_df <- function(DF, .what, .with) {
  table.add.margins <- function(...) addmargins(table(...))
  lapply(DF[.what], table.add.margins, DF[[.with]])
}

Then

# the columns we want to cross tabulate with `Growing`
table_names <- setdiff(names(df1), 'Growing')
df_list <- setNames(list(df1,df2), c('df1','df2'))

lapply(df_list, tabulate_df, .what = table_names, .with = 'Growing')

cross-tabulations on multiple data.frames and columns

1 Answers1