Copilot commented on code in PR #11168:
URL: https://github.com/apache/arrow-rs/pull/11168#discussion_r4091023046
##########
parquet/src/column/page.rs:
##########
@@ -404,6 +404,39 @@ pub trait PageReader: Iterator<Item = Result<Page>> + Send
{
/// column index information
fn skip_next_page(&mut self) -> Result<()>;
+ /// Decodes and returns a dictionary page this reader has previously
+ /// skipped past, if any.
+ ///
+ /// [`Self::skip_next_page`] may skip a dictionary page without decoding
+ /// it, since skipping rows never needs dictionary contents. If decoding
+ /// later reaches a dictionary-encoded data page, the reader is asked for
+ /// the dictionary through this method, which pays the deferred
+ /// decompression exactly once. A chunk skipped end to end never pays it.
Review Comment:
This documentation block is attached to `supports_deferred_dictionary`, but
it currently describes a method that decodes and returns a page through
`take_deferred_dictionary`; the actual method returns only `bool`. Please
rewrite this opening paragraph to document the capability probe, and keep the
decode/return description on `take_deferred_dictionary`, so the public trait
documentation does not promise the wrong return behavior.
--
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]
