Btukfyl

I have a vector with a bunch of criteria that I want to use to cycle through my data frame summarizing my relevant data column.

I reproduce bellow my code, using dplyr and pipe. It works perfectly, so I'll explain my struggle bellow it.

My code:

c1 <- c(0.5,0.5,0.5,1,1,1,2,2,2,2.5,2.5,2,3,3,4,4,4.4,4.5,4.5,5,5.5,6,7,7,8,8.5,9,9.5)

c2 <- c(12,10,40,4,12,7,3,2,1,4,8,10,10,7,7,4,4,4,5,5,6,15,15,25,4,4,7,18)

c3 <- rep(c("AA","BB","CC","DD"), 7)



df <- data.frame(criteria.names = c3, criteria.data = c1, relevant.data = c2,

 stringsAsFactors = FALSE)



user.criteria <- c(0,2,3,5,7,10)





summarised.data <- df %>%

  group_by(criteria.names) %>%

    summarise(class1 = sum(relevant.data[criteria.data >= 0 & criteria.data < 2]),

              class2 = sum(relevant.data[criteria.data >= 2 & criteria.data < 3]),

              class3 = sum(relevant.data[criteria.data >= 3 & criteria.data < 5]),

              class4 = sum(relevant.data[criteria.data >= 5 & criteria.data < 7]),

              class5 = sum(relevant.data[criteria.data >= 7 & criteria.data < 10]))

Here's my expected output:

 summarised.data

# A tibble: 4 x 6

  criteria.names class1 class2 class3 class4 class5

  <chr>           <dbl>  <dbl>  <dbl>  <dbl>  <dbl>

1 AA                 24      1     14      6      4

2 BB                 17      4     11     15      4

3 CC                 40     11     12      0     22

4 DD                  4     12      4      5     43

MY PROBLEM IS: my "user.criteria" vector, whose values I use in my summarization, is gonna come via user input, so there's no guarantee that they are actually going to provide me with the 2,3,5,7,10 values (0 is always going to be there by default) I've explicitly put in my calculations. I've tried using the apply family functions (apply, sapply, lapply, mapply) and adply (plyr package) but so far I haven't been successful in tackling this problem. I'm trying to avoid using explicit loops in R, as the actual database I'm working with is quite huge.

Bellow is an example of my faulty code:

summarised.try <- 1:(length(user.criteria)-1) %>%

  adply(1,function(x){

   df %>%

      group_by(criteria.names) %>%

      summarise(class = sum(relevant.data[criteria.data >=user.criteria[x]

  & criteria.data < user.criteria[x+1]]))})

What I want is to find an elegant way to get the values my user provides me and use them to automatically calculate my summarization, without needing to manually edit my code. Tks!

This function might be the least elegant solution, however it works if we keep same df's column names (i.e. criteria.names, criteria.data, relevant.data ):

library(dplyr)



classifier <- function(criteria, df){



  classified_columns = list()



  for(i in 1:length(criteria) ){



    tmp_class = vector("numeric")



    for( ii in unique(df$criteria.names) ){



      tmp_df = df[df$criteria.names == ii,] 



      if ( i + 1 <= length(criteria)  ){



        tmp_df %>%

          summarise(n = relevant.data[criteria.data >= criteria[i] & criteria.data < criteria[i + 1]] %>% 

                      sum() )  %>%

          .$n %>%

          append(x = tmp_class, values = .) -> tmp_class

      }

    } 



    if( length(tmp_class) > 0 ){



      classified_columns[[paste("class", i, sep = "")]] = tmp_class 

    }       

  }



  data.frame(criteria.names = unique(df$criteria.names),

             as.data.frame(classified_columns)) %>% 

    return(.)

}

Testing function:

classifier(criteria = user.criteria, df = df)

Output:

   criteria.names class1 class2 class3 class4 class5

1             AA     24      1     14      6      4

2             BB     17      4     11     15      4

3             CC     40     11     12      0     22

4             DD      4     12      4      5     43