R Proportion Table Include All N Ways? -

i know done in sas, can't seem figure out how done in r. so, i'm trying create multi-way proportion table of n-ways. mean -

i know how (this made example):

prop_table<-as.data.frame(prop.table(table(df$variable1,df$variable2),1))    view(prop_table)      variablea variableb freq       x1         y1        0.5        x2         y1        0.75       x1         y2        0.5      x2         y2        0.25 

but want one-way proportions...something this:

variablea variableb freq   x1                   0.2   x2                   0.8   x1         y1        0.5   x2         y1        0.75   x1         y2        0.5   x2         y2        0.25 

is there way without separately creating 1-way, 2-way...n-way tables , appending them all?

i don't know of built in function combinations you, it's not many lines of code write own. here's such function

nprop.table <- function(x) {   tbls <- map(function(n) {do.call("table", list(bquote(x[,1:.(n), drop=false]),        dnn=bquote(names(x)[1:.(n)])) )}, ncol(df):1)   props <- map(function(x) as.data.frame(prop.table(x,if(length(dim(x))>1){1} else {numeric(0)})), tbls)   dplyr::bind_rows(props) } 

we can run with

df <- expand.grid(variable1=c("x1","x2"), variable2=c("y1","y2"))[rep(1:4, c(10,60,10,20)),] nprop.table(df) #   variable1 variable2 freq # 1        x1        y1 0.50 # 2        x2        y1 0.75 # 3        x1        y2 0.50 # 4        x2        y2 0.25 # 5        x1      <na> 0.20 # 6        x2      <na> 0.80 

it uses columns of data.frame pass in. if had 3 columns, should still work

nsamp <- function(x, n) sample(x, n, replace=t) df <- data.frame(var1=nsamp(letters[1:3], 50),     var2=nsamp(letters[4:6], 50),     var3=nsamp(letters[7:8], 50)) nprop.table(df)