transforming R function output to match format of Python function output

Question

translating a Python function to R I try to add more data while iterating through values which I split using strsplit() in R. The return value of the original Python function is a list of lists so this should be the same in the R version (vector of vectors). The Python function is

anbaudauer = 12
planzen_list = [
                    'tomaten_6',
                    'karotten_4',
                    'paprika_7',
                    'erdbeeren_5',
                    'koriander_2',
                    'salat_3',
                    'zucchini_4',
                    'gurke_5',
                    'petersilie_2',
                    'radieschen_3'
                ]

def possible_combinations(values, target, with_replacement=True):
    def sub_combinations(index, l, r, t, w):
        if t == sum(filter(lambda i: isinstance(i, int), l)):
            r.append(l)
        elif t < sum(filter(lambda i: isinstance(i, int), l)):
            return
        for u in range(index, len(values)):
            sub_combinations(u if w else (u + 1), l + [values[u].split('_')[0], int(values[u].split('_')[1])], r, t, w)
        return r
    return sub_combinations(0, [], [], target, with_replacement)

raw_combinations = possible_combinations(planzen_list, anbaudauer)

# returns [['tomaten', 6, 'tomaten', 6], ['tomaten', 6, 'karotten', 4, 'koriander', 2], ['tomaten', 6, 'karotten', 4, 'petersilie', 2], ...]

the current state of the R function is

w <- c(
  'tomaten_6',
  'karotten_4',
  'paprika_7',
  'erdbeeren_5',
  'koriander_2',
  'salat_3',
  'zucchini_4',
  'gurke_5',
  'petersilie_2',
  'radieschen_3'
)
n <- length(w)
t <- 12
D <- list()
for (j in 0:n) D[[paste(0, j)]] <- list(c())
for (i in 1:t) D[[paste(i, 0)]] <- list()
for (j in 1:n) {
  for (i in 1:t) {
    D[[paste(i, j)]] <- do.call(c, lapply(0:floor(i/as.numeric(strsplit(w[j], '_')[[1]][2])), function(r) {
      lapply(D[[paste(i-r*as.numeric(strsplit(w[j], '_')[[1]][2]), j-1)]], function(x) c(x, rep(strsplit(w[j], '_')[[1]][1]), r))
    }))
  }
}
D[[paste(t, n)]]

at as.numeric(strsplit(w[j], '_')[[1]][2])I would like to add the name of the vegetable before the number. In essence the R function should return the same as the Python function. Transforming the R functions return value was tricky so far. Basically I tried taking everything apart and stitching it together again. As I haven't used R in a long time I'm sure there is a convenient method to get the same data returned as from the Python function.

`strsplit` is vectorized. You could directly use `strsplit(w, '_')` instead of looping — akrun, Feb 09 '20 at 19:49

GWD · Answer 1 · 2020-02-09T20:36:05.210

1

Why not go with good old RegEx here? For a start:

#first drop the text before the '_' 
w_int <- as.integer(gsub(pattern = ".*_", replacement = "", w))
#repeat the replacement for the numbers at the end to get the vegetable names
w_names <- gsub(pattern = '_[0-9]*$', "", w)
#assign the vegetable names to the Anbaudauer-vector as names 
names(w_int) <- w_names 
w_int
tomaten   karotten    paprika  erdbeeren  koriander      salat ...
     6          4          7          5          2          3 ...
Anbaudauer <- 12

And from there on just check all combinations that add up to Anbaudauer. Potentially something like this ... combinations of numbers to reach a given sum - recursive implementation in R don't want to repeat that here.

edited Feb 09 '20 at 20:36

answered Feb 09 '20 at 19:52

GWD

1,387
10
22

thanks, the issue here is that some plants have the same Anbaudauer (e.g. koriander and petersilie) and that plants can be reused. The suggested implementations gives correctly 103 possible combinations but I want to have the plant names included and not just the numbers that sum to the Anbaudauer as the plants are the distinguishing factor I'm looking for. – Aaron Scheib Feb 10 '20 at 06:51
you can always access vector name attributes via the names() function; additionally the l|s|v*apply family of functions has a USE.NAMES = TRUE|FALSE parameter available, where the names can get preserved - in case you reuse the function in the linked answer; I think there are also package "purrr" functions that help handle this problem – GWD Feb 10 '20 at 11:19
but using ```w_int``` as input for the function does not get the correct result as it is ignoring that e.g. a plant with 2 days is occurring twice. Using ```w_int``` as input for the function results in only 50 combinations instead of the 103. – Aaron Scheib Feb 10 '20 at 16:23
Hi ... have you tried using something like `combn(w_int, min(which(cumsum(w_int)>Anbaudauer), simplify = FALSE)` – GWD Feb 11 '20 at 21:41
I am also not sure how you define correctly in your context - since you were originally asking how to split out names and numbers. The question has changed somewhat or is a different/new one. – GWD Feb 11 '20 at 22:41
BUT if you have all combinations defined as above `comb(...)` it is an exercise of applying something like this `veggies <- sapply(lapply(Combs, function(x) which(cumsum(x) <= target)), names)` to get the veggie names out PLUS removing the duplicates - serch for that on SO – GWD Feb 11 '20 at 22:49

transforming R function output to match format of Python function output

1 Answers1