1

I have a file "ingredients.txt" that has a list of 38 words (one per each line). I have figured out how to read it into an array

getArray() {
    array=() 
    while IFS= read -r line 
    do
        array+=("$line")
    done < "$1"
}

getArray "ingredients.txt"

But now how to get output of all combinations of exactly 4 words from the list of 38?

EDIT: To respond to a comment, I do mean combinations and not permutations, and repetitions are not valid. Four unique words in each combination.

To further clarify, think of it as drawing four marbles of 38 different colors from a bag. No two colors alike. Once you draw a red one, it is impossible to draw another red one for that combination. You draw four, write down the colors, and put them back in the bag, and draw another four. If you get {blue, yellow, purple, red} and {yellow, purple, red, blue} they are not counted separately. I just want combinations, not permutations.

Also, I want each combination printed (going back to the ingredients list): onion, cheese, meat, bun lettuce, carrot, celery, radish carrot, lettuce, celery, vinegar etc.

I hope this is clear.

  • 5
    This strikes me as something you'd probably want to write in a some other programming language, not bash. – Kusalananda Oct 12 '18 at 5:46
  • Please read: Combinations and Permutations And say: Is repetition valid? e.g: (a,a,a,a). Is order important? That is: are (a,b,c,d) and (d,b,a,c) equivalent? – Isaac Oct 12 '18 at 8:09
  • You can also use the bash mapfile command to get the file into an array: mapfile -t array < ingredients.txt – glenn jackman Oct 12 '18 at 14:55
  • I meant what I said, combinations, not permutations. And no, repetition is not valid. Needs to be four unique words. – gracious1 Oct 16 '18 at 23:37
2

In Python:

import itertools.combinations

with open('ingredients.txt') as fd:
    words= fd.readlines()
    for combination in itertools.combinations(words, 4):
        print(combination)
| improve this answer | |
  • This can be inserted into a bash script as is? (Not experienced with python.) – gracious1 Oct 17 '18 at 4:31
  • I see that you can't, what is the best syntax for inserting this into bash script? – gracious1 Oct 17 '18 at 4:37
0

I tried by using python

>>> a
['praveen', 'ajay', 'abhi', 'chetan', 'abh', 'cat']


for i in range(0,len(a),4):
...     print a[i:i+4]
...
['praveen', 'ajay', 'abhi', 'chetan']
['abh', 'cat']
| improve this answer | |
0

Just for the fun of it, a version in plain bash (which I do not recommend using for long lists, it is quite slow (as should be expected), it takes more than two minutes to do in shell what in c gets done in half a second.):

#!/bin/bash -
set -u

readarray -t b <ingredients.txt

r=${1:-3}
n=${2:-5}

main(){
          combinations
      }


combinations(){  : ${r:=3}    ${n:=5}    # default values

        ## The first line of elements is easy, fill an array of r elements:
        for ((i=0 ; i<r ; i++)); do 
            a[i]=$i
        done
        printelements

        ## Up to the maximum permitted value of the first array element.
        while (( a[0] < n-r )); do
            ## search backwards on the array for a value less than maximum.
            for ((i = r-1; i >= 0; i--)); do
            ## If an item is below its maximum permitted value...
            if ((a[i] < n-r+i )); then 
                ## increment it:
                ((a[i]++))
                break
            fi
            done
            ## Fill the rest of the array with consecutive values:
            for (( i = i + 1 ; i < r ; i++ )); do
            (( a[i] = a[i-1] + 1 ))
            done
            ## Print the current combination of items:
            printelements
        done
         }

printelements(){ : #p=${a[@]/%/ }; printf '%s\n' "<${p% }>"; }
                 s=""
                 for i in "${a[@]}"; do
                     printf '%s' "$s" "${b[i]}"
                     s=" "
                 done
                 echo
               }

main

Running it as:

$ ./script 4 38

Will print:

$ ./script 4 38 | wc -l
73815

As math will confirm (fact is a library defined function):

$ bc <<<"r=4;n=38;fact(n)/(fact(r)*fact(n-r))"    # n! / ( r! × (n-r)! )
73815
| improve this answer | |
-1

Well, try with below code

getArray() {
    array=() 
    while IFS= read -r line 
    do
        array+=("$line")
    done < "$1"
    for i in ${!array[@]}
    do
           if [ ${#array[$i]} == 4 ]; then
                    echo "${array[$i]}"
           fi
    done
}

getArray "ingredients.txt"
| improve this answer | |
  • There's readarray in bash, and the code does not give any combinations of four words, only the words that are four characters long. This you could have done through a fairly simple grep command or similar, or awk 'length == 4'. – Kusalananda Oct 12 '18 at 6:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.