Aw: Re: Building a binary vector out of dichotomous variables

Hi Tom,

thanks for your reply.

Yes, that's exactly what I am looking for. I did not know about the automatic type conversion in R.

#-- cut --
ds_example <-
  structure(
    list(
      year2013 = c(0, 0, 0, 1, 1, 1, 1, 0),
      year2014 = c(0,
                   0, 1, 1, 0, 0, 1, 1),
      year2015 = c(0, 1, 1, 1, 0, 1, 0, 0)
    ),
    .Names = c("year2013",
               "year2014", "year2015"),
    row.names = c(NA, 8L),
    class = "data.frame"
  )

#-- Proposal: works!
as.numeric(with(ds_example,paste(1,year2013,year2014,year2015,sep='')))

# I store my know-how about R in functions for later use.

#--? Putting it in a function - does not work!
t_make_binary_vector <- function(dataset,
                                 input_variables,
                                 output_variable = "binary_vector") {
  dataset[output_variable] <- "1"
  print(dataset[output_variable])
  
  for (variable in input_variables) {
    print(variable)
    dataset[output_variable] <- paste(dataset[output_variable],
                                      dataset[variable], 
                                      sep='')
  }
  
  # print(dataset[output_variable])

  dataset[output_variable] <- as.integer(dataset[output_variable])
  
  return(dataset)
}

t_make_binary_vector(dataset = ds_example,
                     input_variables = c("year2013", "year2014", "year2015"),
                     output_variable = "binary_vector")


#-- Doesn't work either.
t_make_binary_vector <- function(dataset,
                                 input_variables,
                                 output_variable = "binary_vector") {
  dataset[output_variable] <- as.integer(paste(1, dataset[ , input_variables], sep = ''))

  return(dataset)
}

t_make_binary_vector(dataset = ds_example,
                     input_variables = c("year2013", "year2014", "year2015"),
                     output_variable = "binary_vector")

#-- cut --

Why is R taking the parameter value itself to paste it together instead of referencing the variable within the dataset?

What did I get wrong about R? How can I fix it?

Kind regards

Georg

Gesendet: Donnerstag, 16. Juni 2016 um 16:13 Uhr
Von: "Tom Wright" <tom at maladmin.com>
An: G.Maubach at weinwolf.de
Cc: "R. Help" <r-help at r-project.org>
Betreff: Re: [R] Building a binary vector out of dichotomous variables

Does this do what you want?

as.numeric(with(ds_example,paste(1,year2013,year2014,year2015,sep='')))

On Thu, Jun 16, 2016 at 8:57 AM,  <G.Maubach at weinwolf.de> wrote:

Hi All,

I need to build a binary vector made of a set of dichotomous variables.

What I have so far is:

-- cut --

ds_example <-
  structure(
    list(
      year2013 = c(0, 0, 0, 1, 1, 1, 1, 0),
      year2014 = c(0,
                   0, 1, 1, 0, 0, 1, 1),
      year2015 = c(0, 1, 1, 1, 0, 1, 0, 0)
    ),
    .Names = c("year2013",
               "year2014", "year2015"),
    row.names = c(NA, 8L),
    class = "data.frame"
  )

attach(ds_example)
base <- 1000
binary_vector <- base + year2013 * 100 + year2014 * 10 + year2015
detach(ds_example)

binary_vector

ds_example <- cbind(ds_example, binary_vector)

varlist <- c("year2013", "year2014", "year2015")

base <- 10^length(varlist)

binary_vector <- NULL

for (i in 1:3) {
  binary_vector <-
   base +
   ds_example [[varlist[i]]] * base / (10 ^ i)
}

ds_example <- cbind(ds_example, binary_vector)

message("Wrong result!")
ds_example

-- cut --

How do I get vectors like  1000 1001 1011 1111 1100 1101 1110 1010 for
each case?

Is there a better approach than mine?

Kind regards

Georg