There are copies, and there are counterfeits. The copies may well be reasonable substitutes for the originals; the counterfeits may be entirely unusable.
I've examined and noted the differences between an original Hitachi B649 and a few possibly copied D669As - the most obvious is the shape of the metal heatsink tab at the back. The front of the copied 669As look similar to the pictures at the beginning of this thread. The original Hitachi always has 3 'rays' around the mounting hole and a very clean-cut outline of the metal tab at the back, the copies occasionally have no rays and the metal tab outline is rough. Incidentally, some of the possible copies came from a sealed pack with the Hitachi logo and the text Taiwan at the bottom - these have 3 rays like the original Hitachi.
The DC beta on most of the probable 669A copies measures around 140-170, which is a tad higher than the tight cluster around 140 or so for the real Hitachi B649. This is probably not conclusive, because of normal NPN/PNP variations, process variations, binning, batches, etc.
However, if the beta is clustered higher than 200, it warrants careful examination. I have one very different-looking D669A outlier with a beta of 360 or so - this has a mounting hole with no 'ray' at all, and a very rough finish on the leads. This is most probably a counterfeit - a small-signal transistor masquerading as a D669A.