Abdallah-Afifi commented on issue #11985: URL: https://github.com/apache/arrow-rs/issues/11985#issuecomment-6554940212
That matches what I see in the code. `encode_vector` derives `bit_width` from the FOR range over the surviving non-exceptions and then writes the packed section whenever `bit_width > 0`, with no comparison against what those values would cost as exceptions instead. The placeholder fill keeps the range tight when the non-exceptions are clustered, but nothing helps when two of them sit far apart. The decode side already looks ready for it: `bit_width == 0` fills the vector with the frame of reference, and the header validation only rejects `num_exceptions > num_elements`, so a fully demoted vector round-trips without a format or reader change. Are you planning to pick this up? Happy to leave it with you, since you have real datasets to check it against — otherwise I can take it. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
