Thank you very much Daniel and Chris. As soon as my code finish running, I will try your solutions.

Amazing to have this kind of help.

Regards Rodrigo

Yes it very simple to make it 10000 x faster when testing a 12 million strings list ;)

The simplest solution is maybe to build first an optimized (Dispatch) list of rules (Rule) which associates the strings with their position in the database.

Let's make a 12 million strings list similar to the one (skustofind) you give in your example:

chars = Join[CharacterRange["A", "Z"], CharacterRange["0", "9"]];
skustofind = Table[RandomChoice[chars, 10] // StringJoin, 12000000];

Then we build the optimized table describing which string corresponds to which position:

dispatch = Dispatch@Table[ skustofind[[i]] -> i, {i, 1, Length@skustofind}];

(Or in a more functional programming way you could write also dispatch=Dispatch@Thread@Rule[skustofind, Range[Length@skustofind]])

That's it, let's compare the time it takes using Position and the Dispatch table.

For example if the test string is:

testString=skustofind[[10000000]]

EH2AMN0CW0

then

Position[skustofind, testString] // AbsoluteTiming

{0.268778, {{10000000}}}

while

testString /. dispatch // AbsoluteTiming

{0.00001, 10000000}

The dispatch approach is about (0.26/0.00002 =)10000x faster here. The bigger your string list is, the longer it will take for Position to find while it is practically the same constant time with the dispatch way.

Alternative approaches (to build the same kind of optimized table as dispatch) is for example to use Association or a DataStructure but the final timings are the same.