2
$\begingroup$

I have two lists:

list1={4,6}
list2={1,2}

and I now want a function f that produces

f[list1_,list2_]:={1,1,1,1,2,2,2,2,2,2}

so 4 times the 1 and 6 times the 2. f should be generic so that

3list1={2,3,7}; 3list2={1,2,3}

yields

res={1,1,2,2,2,3,3,3,3,3,3,3}

If you know several ways to do this please provide the fastest one. thanks a lot.

$\endgroup$
6
  • $\begingroup$ Table[list2[[#]], {list1[[#]]}] & /@ Range@Length@list1 // Flatten $\endgroup$
    – eldo
    Commented Jun 3, 2014 at 18:16
  • $\begingroup$ Flatten @ MapThread[ ConstantArray[ #2 , #1 ] &, { list1 , list2 } ] $\endgroup$
    – mfvonh
    Commented Jun 3, 2014 at 18:20
  • $\begingroup$ Flatten@MapThread[ConstantArray[##] &, {list2, list1}] $\endgroup$
    – eldo
    Commented Jun 3, 2014 at 18:24
  • $\begingroup$ Related: (40724) and related Stack Overflow question: (763915) $\endgroup$
    – Mr.Wizard
    Commented Jun 3, 2014 at 21:19
  • $\begingroup$ @Mr.Wizard - related or not - how would you write this ? $\endgroup$
    – eldo
    Commented Jun 3, 2014 at 22:30

2 Answers 2

4
$\begingroup$
f[list1_, list2_] := Inner[Table[#2, {#1}] &, list1, list2, Join];
$\endgroup$
1
  • $\begingroup$ this is very clear and understandable, thanks. $\endgroup$
    – spore234
    Commented Jun 3, 2014 at 19:06
3
$\begingroup$
list1 = {4, 6}
list2 = {1, 2}

ClearAll[f];
f = Join @@ MapThread[Table, {#2, List /@ #1}] &

f[list1, list2]
(* {1, 1, 1, 1, 2, 2, 2, 2, 2, 2} *)

f[{2, 3, 7}, {1, 2, 3}]
(* {1, 1, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3} *)

or

h = Join @@ (Table @@@ Transpose[{#2, List /@ #1}]) &
h[{2, 3, 7}, {1, 2, 3}]
(* {1, 1, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3} *)

or

g = Join @@ (ConstantArray @@@ Transpose[{#2, #1}])&

g2 = Inner[ConstantArray, #2, #1, Join] &

Timings:

bh = Inner[Table[#2, {#1}] &, #1, #2, Join] &; (* BobHanlon *)

{listA, listB} = RandomInteger[100, {2, 100000}];
(rs1 = bh[listA, listB]); // Timing
(* {0.265625, Null} *)
(rs2 = g[listA, listB]); // Timing
(* {0.062500, Null} *)
(rs3 = g2[listA, listB]); // Timing
(* {0.078125, Null} *)
(rs4 = h[listA, listB]); // Timing
(* {0.296875, Null} *)
(rs5 = f[listA, listB]); // Timing
(* {0.296875, Null} *)
rs1 == rs2 == rs3 == rs4 == rs5
(* True *)
$\endgroup$
5
  • $\begingroup$ your solution is the fastest so far. $\endgroup$
    – eldo
    Commented Jun 3, 2014 at 18:37
  • $\begingroup$ @eldo As usual I'm the slowest so far :) $\endgroup$
    – mfvonh
    Commented Jun 3, 2014 at 18:46
  • 1
    $\begingroup$ With two 10^6 - Integer-Lists, MM 9.01 and 4 cores: Bob Hanlon -> 1.34, kguler -> 1.53, mfvonh -> 2.07, eldo -> 3.15 (up to 3 repetitions). $\endgroup$
    – eldo
    Commented Jun 3, 2014 at 18:58
  • $\begingroup$ @kguler - why is "g" so much faster than "f" given the fact that both are kguler-offsprings :) $\endgroup$
    – eldo
    Commented Jun 3, 2014 at 19:31
  • 1
    $\begingroup$ @eldo, it is because i could not get g working properly first time I tried and f worked right away :) I cannot locate the source on this site right now, but Join@@.. is usually faster than Flatten and ConstantArray is also faster than alternatives like Table and Array. $\endgroup$
    – kglr
    Commented Jun 3, 2014 at 19:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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