Do learned SA and Hatachi signals have the same conflict?

Both SA and Hitachi are consistent enough across models, that I'd be surprised at any significant difference between well learned codes and those from the database.

If learned codes conflict, you can look at their Pronto Hex and try to guess why they conflict. I doubt that you'll see why because SA and Hitachi codes don't have enough similarity for a conflict to be reasonable.

But in most well documented cases are IR code conflict, the conflict wasn't reasonable (excluding of course cases where devices used the same codes). The wrong device
would receive a signal with very little similarity to any signal it should understand and would simply decode it as one of its own commands for no good reason.