Nonparametric alternative for Spearman's rho for correlation strength?

My question is as simple as it may be difficult: is there any other nonparametric test besides Spearman's rho to compute the correlation strength between a predictor and an outcome variable of which the first is almost certain to have a skewed distribution?

The reason why I'm asking is that in MySQL, the computer database language that I'm tied to, it is very difficult, if at all possible, to rank my values because of many duplicate values. And to rank 'em manually is not doable either, because there will be between two and six thousand entries.

So I'm actually looking for a nonparametric correlation strength test with an easier ranking method, assuming that ranking is a key principle of all of them.
Last edited:
Kendall's tau can be computed without ranking, although the computation is O(N^2). You should be aware that in situations with many ties, the null distribution changes considerably.
I had looked at Kendal's tau before, but my descriptive text book clearly gives the impression that it is a test for paired observations, for a degree of agreement/similarity. However, your post made me look it up on Wiki, which page shows a lot more possibilities, especially Tau-b. I still have to turn that into (a) workable MySQL formula(s), but that would seem quite doable. Thanks! :)