(2): I do it once in a while.
I load "tags.csv" on Excel, filter all the OK entries and all the @nnc separately, and import both index lists to matlab. There:
> arrayOK = false(1, 3000000);
> arrayOK(filterOK) = true;
> arrayNNC = false(1, 3000000);
> arrayNNC(filterNNC) = true;
>
> seq = 1:3000000;
> seq = seq(filterNNC & filterOK);

and that's it ^^
If you want, I can put a link to a results file here on the Wall tomorrow.

> so that they can remove both tags
Why would we want to remove an "OK" tag?

> remove both tags

"OK" tags should never be removed.
Their most important function ist to say: Even if the author isn't a native speaker, the sentence has been checked by a native speaker and therefore you can trust it as if it were owned by a native speaker.

Of course there are a lot of other sentences tagged with "OK" too, but it's just a sentence owned by a non-native speaker where this tag makes the really great difference in the eyes of the user.

Of course, if a sentence has both "@needs native check" and "OK", than "@needs native check" makes no sence any longer and should be removed. Mostly this is done right away when the sentence has been checked and tagged "OK".

Here's the list I talked about. There were only 22 results:

20337
23054
59378
65213
71100
145969
320988
326866
450824
1065935
1443415
1523131
1763683
2069475
2120582
2180411
2180700
2220721
2220744
2223944
2265117
2468564

you can verify both tags until a maintainer removes the respective @nnc