My suggestion.For which tool/technology or hadoop or ..... Maybe you can put under the right forum, someone who knows better about it can answer.
Do you mean a column defined with GENERATED ALWAYS AS IDENTITY?
This is the the way it is implemented in TD, it's not one sequence, it's a parallel sequence. Each AMP/PE reserverves a batch of 100,000 numbers and assignes them sequentially. The batch size can only be modified on a system level, but will never be set to 1.
If you need sequential values without gaps you should do it within SQL using ROW_NUMBER.