Deleting duplicates which are identified as equal up to rearrangement of sublists

Question

I have a list of integer 2-tuples s = { {{3,4},{11,5}} , {{11,5},{3,4}}, {{5,5},{3,2}} } (in reality this set is much larger).

I am trying to delete duplicates from this list. Specifically, I want to treat lists within s which, up to a reordering, contain the same lists (2-tuples) as the same, and keep only one.

For instance, if s were my actual list, I would want to convert it into the following list: { {{3,4},{11,5}} , {{5,5},{3,2}} }.

I can't figure out an efficient way of doing this; in reality, my list s is quite large, and its sublists also consist of many elements.

Answer 1

DeleteDuplicates[Sort /@ s]

{{{3, 4}, {11, 5}}, {{3, 2}, {5, 5}}}

You can also use Gather or GatherBy:

Gather[Sort /@ s][[All, 1]]

{{{3, 4}, {11, 5}}, {{3, 2}, {5, 5}}}

GatherBy[s, Sort][[All, 1]]

{{{3, 4}, {11, 5}}, {{5, 5}, {3, 2}}}

All three above are faster than DeleteDuplicatesBy[Sort] for long lists of lists:

SeedRandom[1]
ss = RandomInteger[10, {100000, 2, 2}];

r1 = DeleteDuplicates[Sort /@ ss]; // RepeatedTiming // First

0.038

r2 = Gather[Sort /@ ss][[All, 1]]; // RepeatedTiming // First

0.12

r3 = GatherBy[ss, Sort][[All, 1]]; // RepeatedTiming // First

0.13

r4 = DeleteDuplicatesBy[Sort]@ss; // RepeatedTiming // First

0.15

Sort /@ r1 == Sort /@ r2 == Sort /@ r3 == Sort /@ r4

True

SeedRandom[1]
ss = RandomInteger[10, {100000, 50, 5}];

r1 = DeleteDuplicates[Sort /@ ss]; // RepeatedTiming // First

1.26

r2 = Gather[Sort /@ ss][[All, 1]]; // RepeatedTiming // First

1.69

r3 = GatherBy[ss, Sort][[All, 1]]; // RepeatedTiming // First

1.61

r4 = DeleteDuplicatesBy[Sort]@ss; // RepeatedTiming // First

3.27

Sort /@ r1 == Sort /@ r2 == Sort /@ r3 == Sort /@ r4

True