(arrow-rs) branch main updated: avoid redundant parsing of repeated value in RleDecoder (#6834)

[tustvold]

This is an automated email from the ASF dual-hosted git repository.

tustvold pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/arrow-rs.git


The following commit(s) were added to refs/heads/main by this push:
     new fa6d5e171b avoid redundant parsing of repeated value in RleDecoder 
(#6834)
fa6d5e171b is described below

commit fa6d5e171b12cf7cc20fe64744afcc5e7f28b82e
Author: Jinpeng <zjp...@163.com>
AuthorDate: Thu Dec 5 05:41:54 2024 -0500

    avoid redundant parsing of repeated value in RleDecoder (#6834)
    
    Co-authored-by: jp0317 <zjp...@gmail.com>
---
 parquet/src/encodings/rle.rs | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/parquet/src/encodings/rle.rs b/parquet/src/encodings/rle.rs
index 0c708c1265..d089ba7836 100644
--- a/parquet/src/encodings/rle.rs
+++ b/parquet/src/encodings/rle.rs
@@ -369,17 +369,17 @@ impl RleDecoder {
     }
 
     #[inline(never)]
-    pub fn get_batch<T: FromBytes>(&mut self, buffer: &mut [T]) -> 
Result<usize> {
+    pub fn get_batch<T: FromBytes + Clone>(&mut self, buffer: &mut [T]) -> 
Result<usize> {
         assert!(size_of::<T>() <= 8);
 
         let mut values_read = 0;
         while values_read < buffer.len() {
             if self.rle_left > 0 {
                 let num_values = cmp::min(buffer.len() - values_read, 
self.rle_left as usize);
+                let repeated_value =
+                    
T::try_from_le_slice(&self.current_value.as_mut().unwrap().to_ne_bytes())?;
                 for i in 0..num_values {
-                    let repeated_value =
-                        
T::try_from_le_slice(&self.current_value.as_mut().unwrap().to_ne_bytes())?;
-                    buffer[values_read + i] = repeated_value;
+                    buffer[values_read + i] = repeated_value.clone();
                 }
                 self.rle_left -= num_values as u32;
                 values_read += num_values;