[RS] Contains vs. position function

[gartmin]

I do a lot of filtering on character values among massive data sets (millions of records). I know I can filter records with contains {[var1] contains 'Character string'} or with a position statement {position('charactervalue', [var1]) > 0}

Any idea on which one is more efficient? Is there a better way?

[COGNOiSe administrator]

I'd have to say contains should be faster, sice the second approach requires two operations, but it will depend on the implementation of either.

How about LIKE? Or having a separate dictionary table, for each letter, all GUIDs of rows where it can be found?