Hi, I had a question regarding bool column stream in a stripe, is it mandatory to have stream(compressed data) to be continuous between two row groups?
bool columns are compressed as bits, so there are chances where a byte is shared between two row groups if column has nulls, but writer can choose to start with a fresh byte and leave the partial bits as it is and update Row Group Index to signify where to start. I didn't find any explicit expectation for stream to be continuous, it might be implicit, so just wanted to confirm about it. With Regards, Ramakrishna Prabhu
