jhorstmann commented on code in PR #8031:
URL: https://github.com/apache/arrow-rs/pull/8031#discussion_r2247919728
##########
parquet-variant/src/builder.rs:
##########
@@ -64,19 +64,55 @@ fn write_offset(buf: &mut Vec<u8>, value: usize, nbytes:
u8) {
buf.extend_from_slice(&bytes[..nbytes as usize]);
}
-/// Write little-endian integer to buffer at a specific position
-fn write_offset_at_pos(buf: &mut [u8], start_pos: usize, value: usize, nbytes:
u8) {
- let bytes = value.to_le_bytes();
- buf[start_pos..start_pos + nbytes as
usize].copy_from_slice(&bytes[..nbytes as usize]);
-}
-
/// Append `value_size` bytes of given `value` into `dest`.
fn append_packed_u32(dest: &mut Vec<u8>, value: u32, value_size: usize) {
let n = dest.len() + value_size;
dest.extend(value.to_le_bytes());
dest.truncate(n);
}
+/// An iterator that yields the bytes of a packed u32 iterator.
+/// Will yield the first `packed_bytes` bytes of each item in the iterator.
+struct PackedU32Iterator<T: Iterator<Item = [u8; 4]>> {
+ packed_bytes: usize,
Review Comment:
I haven't looked into the variant implementation or this PR in detail, but
does this iterator need a custom implementation or can it be an anonymous `impl
Iterator<Item=u8>`? It looks like the implementation is basically
`iter.flat_map(u32::to_le_bytes)`. The benefit of that is that if the
underlying iterator has a trusted length then the flat-mapped iterator is still
trusted, and extending a slice from a `TrustedLen` iterator is much more
efficient.
--
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: github-unsubscr...@arrow.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
