Searching for a string in a very efficient way

2 views (last 30 days)
Hi.
I have two cell arrays, A and B. A contains 3000 strings and B contains 107,000 strings. Some of the strings in B might also be present in A. Each of these strings is a 7 letter word. I need to find which strings in B are also present in A. I did this by writing the following:
[~,p] = ismember(A, B);
p(p==0) = []; % removing zeros from the p array
I then use the indices in p to get the strings in B that are also in A. This works but takes a long time to compute (0.4 sec). Since I need to do this step about 30 times (for other A and B), the overall time becomes large. Does anyone have an idea of how to make this string search faster. With thanks in advance, G

Accepted Answer

Azzi Abdelmalek
Azzi Abdelmalek on 17 Jul 2013
Edited: Azzi Abdelmalek on 17 Jul 2013
Instead of ismember(B, A), You should use
ismember(A,B);
  1 Comment
gwar
gwar on 17 Jul 2013
In my case ismember(A,B) or ismember(B,A) have very similar computation times. I have edited my question since I mixed up the A and B.

Sign in to comment.

More Answers (1)

Jan
Jan on 17 Jul 2013
ismember sorts the input strings to allow for an efficient binary search. If you call this 30 times, sorting the larger cell string explicitly will save 29 sorting procedures.
  1 Comment
gwar
gwar on 17 Jul 2013
You mean combining all the 30 A´s and 30 B´s and calling ismember only once on the combined cell array?

Sign in to comment.

Categories

Find more on Characters and Strings in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!