etseidl commented on code in PR #8729:
URL: https://github.com/apache/arrow-rs/pull/8729#discussion_r2473598947
##########
parquet/src/encodings/rle.rs:
##########
@@ -459,7 +469,10 @@ impl RleDecoder {
self.rle_left -= num_values as u32;
values_read += num_values;
} else if self.bit_packed_left > 0 {
- let bit_reader = self.bit_reader.as_mut().expect("bit_reader
should be set");
+ let bit_reader = self
+ .bit_reader
+ .as_mut()
+ .ok_or_else(|| general_err!("bit_reader should be set"))?;
Review Comment:
```suggestion
let Some(bit_reader) = self.bit_reader.as_mut() else {
return Err(general_err!("bit_reader should be set"));
};
```
is a bit more concise.
##########
parquet/src/arrow/record_reader/definition_levels.rs:
##########
@@ -131,9 +131,12 @@ impl DefinitionLevelBufferDecoder {
impl ColumnLevelDecoder for DefinitionLevelBufferDecoder {
type Buffer = DefinitionLevelBuffer;
- fn set_data(&mut self, encoding: Encoding, data: Bytes) {
+ fn set_data(&mut self, encoding: Encoding, data: Bytes) -> Result<()> {
match &mut self.decoder {
- MaybePacked::Packed(d) => d.set_data(encoding, data),
+ MaybePacked::Packed(d) => {
+ d.set_data(encoding, data);
+ Ok(())
+ }
MaybePacked::Fallback(d) => d.set_data(encoding, data),
}
Review Comment:
```suggestion
Ok(match &mut self.decoder {
MaybePacked::Packed(d) => d.set_data(encoding, data),
MaybePacked::Fallback(d) => d.set_data(encoding, data)?,
})
```
##########
parquet/src/encodings/rle.rs:
##########
@@ -321,14 +321,15 @@ impl RleDecoder {
}
#[inline]
- pub fn set_data(&mut self, data: Bytes) {
+ pub fn set_data(&mut self, data: Bytes) -> Result<()> {
if let Some(ref mut bit_reader) = self.bit_reader {
bit_reader.reset(data);
} else {
self.bit_reader = Some(BitReader::new(data));
}
- let _ = self.reload();
+ self.reload()?;
Review Comment:
I still want to convince myself that ignoring the result is ok here, but
this isn't a behavior change.
--
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]
