Find unique elements based on a given key

Question

Input

Take a list of values x_i each paired with a key y_i.

[(x1, y1), (x2, y2), ...]

Output

Return a list L containing only values from the set {x_i}.

• The length of L must be equal to the number of unique keys k in the set {y_i}.
• For each unique key k there must be a value from {x_i} that has key k.

Details

• Standard loopholes disallowed.
• You can assume all values in the input will be nonnegative integers.
• There may be duplicate values and keys.
• You can assume there is at least one value/key pair in the input.
• If you prefer to take two lists of equal length as input (one for values, one for keys) that is fine.
• You may not take any other input.
• The order of the list you output does not matter.
• The x_i you choose for each key does not matter.

For example, with input [[0, 0], [1, 3], [2, 3]] you can return either [0, 1] or [0, 2] or any permutation of these.

Examples

[[1, 2], [3, 2], [3, 0]]  -> [1, 3] or [3, 3]
[[7, 2], [7, 0], [7, 1]]  -> [7, 7, 7]
[[4, 0], [4, 0], [9, 1], [5, 2]]  -> [4, 9, 5]
[[9, 1], [99, 10], [5, 5], [0, 3]]  -> [9, 99, 5, 0]

Fewest bytes wins.

Answer 1

Python 2, 34 bytes

lambda a,b:dict(zip(b,a)).values()

Try it online!

Takes input as list of values and list of keys.
Generate dictionary, swapping keys and values, which leaves only unique y values. Return all corresponding x values

Answer 2

Jelly, 2 bytes

Ḣƙ

Try it online!

Takes two lists of equal length, first is keys, second is values.

Answer 3

J, 4 bytes

{./.

How?

The left argument x is a list of key, the right one y- a list of values

/. groups y according x

{. takes the first element of each group

Try it online!

Answer 4

Clojure, 20 18 bytes

(comp vals zipmap)

This takes lists of values and keys as arguments, in that order.

Answer 5

Python 3, 42 bytes

lambda a,b:[dict(zip(b,a))[q]for q in{*b}]

Try it online!

a = values
b = keys

Answer 6

Haskell, 49 47 bytes

map fst.nubBy((.snd).(==).snd)
import Data.List

Try it online! Input as a list of tuples, e.g. [(1, 2), (3, 2), (3, 0)].

---

Input as list of lists (49 bytes)

map(!!0).nubBy(\a b->a!!1==b!!1)
import Data.List

Try it online!

Answer 7

JavaScript (ES8), 43 bytes

Takes input as 2 distinct lists in currying syntax (values)(keys).

v=>k=>v.filter(o=(x,i)=>o[j=k[i]]?0:o[j]=1)

Try it online!

Answer 8

Husk, 4 bytes

ṁhüt

Try it online!

How?

ṁhüt – Full program. 
  üt – Group by tail (the lists without the first element).
ṁh   – Map heads (the lists without the last element) and concatenate the results.

Answer 9

Stax, 5 bytes

╠HB%§

Run and debug it

Takes two arrays, first values, then keys.

Explanation:

cu:I@J Full program, unpacked
cu     Push a unique copy of the keys.
  :I   Indices of the unique elements
    @  Index the values array
     J Join by space
       Implicit output

Answer 10

R, 30 bytes

function(x,y)x[!duplicated(y)]

Try it online!

Answer 11

Japt, 8 bytes

â £VgUbX

Japt Interpreter

Thanks to Shaggy for saving 1 last byte

Completely reworked the logic. Takes some cues from Luis' answer, but I think it's still improved. Now takes input as two lists, keys, values. Apparently I'm still short of optimal though.

Explanation:

â £         For each unique Key X
     Ub     find the first index in "keys"
       X    which is equal to X
   Vg       then take the "value" with the same index

Answer 12

Ruby, 27 bytes

->a,b{b.zip(a).to_h.values}

Try it online!

Takes input as two arrays (the footer transforms the original test cases into this format).

Answer 13

05AB1E, 4 bytes

DÙkè

Take two input lists: first the values, then the keys.

Try it online or verify all test cases.

Explanation:

D        # Duplicate the values-list
 Ù       # Get all unique items of the values-list
         #  i.e. [0,0,1,0,2] → ['0','1','2']
  k      # Get all (first) indices of these unique values
         #  i.e. [0,0,1,0,2] and ['0','1','2'] → [0,2,4]
   è     # And use this index to get the key from the keys-list
         #  i.e. [0,2,4] and [4,4,9,5,4] → [4,9,4]

Answer 14

Java 8, 82 bytes

A lambda from a stream of int[] pairs to java.util.Collection<Integer>.

m->m.collect(java.util.stream.Collectors.toMap(a->a[1],a->a[0],(a,b)->a)).values()

Try It Online

Answer 15

Julia 0.6, 36 34 32 bytes

A->values(Dict(y=>x for(x,y)=A))

Try it online!

(shaved off two bytes thanks to @JonathanFrech)
(another two bytes by replacing ∈ with = in comprehension)

The input format specified in the question [[1, 2], [2, 7]] works as it is in Julia as an array of arrays containing (potential) key value pairs, the only thing to take care of is that the key comes second and value first.

---

Slight change, for the same bytecount,

A->values(Dict((y,x)for(x,y)∈A))

Try it online!

Answer 16

Julia 0.6, 29 26 19 bytes

values∘Dict∘zip

Try it online!

Point-free style. Takes input as an array of keys and an array of values.

---

Older solution:

29 26 bytes

v\k=values(Dict(zip(k,v)))

Try it online!

-3 bytes using operator syntax instead of lambda

Takes input as an array of values and an array of keys.

Answer 17

MATL, 9 bytes

viiQ(5Mu)

Try it online!

Probably suboptimal, but hey, it works!

Uses the facts that (a) input is guaranteed to have only non-negative integers, (b) MATL extends an array of you try to assign to an index that doesn't exist.

v - create an empty array in the stack

i - get array of values

iQ - get array of keys, increment by 1 (so the minimum value is 1, not 0, since MATL indexing is 1-based)

( - assignment indexing - use the array of keys as indices and assign the values to those indices (if any keys are repeated, only the last value remains in that location)

5M - gets the last input of the last call - which would be the array of indices we used

u) - take a unique list of those indices, index with that list, and leave that result (which is a list of values of unique keys) in the stack

Answer 18

Japt, 24 20 16 8 bytes

-4 bytes thanks to @Shaggy

Takes input as key, values

â £VgUbX

Try it online!

---

Japt, 20 18 bytes

r@bY <Z?X:XpVgY}[]

Try it online!

Answer 19

Pyth, 7 bytes

hMhM.ge

All testcases.

Answer 20

Perl 5 -pa, 24 bytes

%a=@F;$_=join$/,values%a

Try it online!

Takes input in the format key value key value key value .... TIO footer is only to separate test cases.

Answer 21

Perl 6, 25 bytes

{.unique(:as(*[1]))[*;0]}

Try it

Expanded:

{  # bare block lambda with implicit parameter $_

  .unique(  # find the unique results (implicit method call on $_)

    :as(    # use this code to compare
      *[1]  # the second value (WhateverCode lambda)
    )

  )[        # index into that
    *;      # all in the top level
    0       # only the first value from each
  ]
}

Answer 22

C-Sharp, 63 bytes

object _(int[][]p)=>p.GroupBy(i=>i[1]).Select(g=>g.First()[0]);

Returns an enumerable of integers.

Answer 23

Wolfram Language (Mathematica), 25 bytes

#&@@#&@@@#~GatherBy~Last&

Try it online!

Answer 24

Rust, 81 bytes

|a,b|b.zip(a).collect::<std::collections::HashMap<_,_>>().into_iter().map(|v|v.1)

Try it online!

Takes two iterators, returns an iterator.

Answer 25

Perl 6, 12 bytes

{%$_.values}

Try it online!

Coerces the given list to a Hash and then returns the values. This works on both a list of pairs, and a list of key, value, key, value....