bharadwaj-pendyala commented on issue #11018: URL: https://github.com/apache/arrow-rs/issues/11018#issuecomment-5585843350
`skip` bails out at `decoding.rs:860` when `values_per_mini_block` isn't 32 or 64; Photon's 256 hits the `_` arm. `get` handles the same page fine by chunking against the caller's buffer, so this only affects skipping. For that TODO, I'd loop over the miniblock using the current 32/64 buffer instead of sizing it off `values_per_mini_block` and eating the allocation. That keeps the common path allocating exactly what it does today. The buffer only advances `last_value`; `BitReader::get_batch` tracks its own bit offset, so reading 256 values as four calls of 64 leaves the reader where one call of 256 would. Then `non_standard_delta_blocks` in `bad_data.rs` can replace the "cannot skip miniblock of size 128" assertion with the `skip should succeed` one already commented out above 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]
