From 5bf9b64a24d2d8ae2af2f967bc2c70544f7f5b9e Mon Sep 17 00:00:00 2001
From: Xiangpeng Hao <haoxiangpeng123@gmail.com>
Date: Wed, 12 Jun 2024 18:09:42 -0400
Subject: [PATCH 1/2] avoid copy/allocation when build from offset buffer

---
 parquet/src/arrow/buffer/offset_buffer.rs | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)

diff --git a/parquet/src/arrow/buffer/offset_buffer.rs b/parquet/src/arrow/buffer/offset_buffer.rs
index 32bb9d0862b9..89484181d0cc 100644
--- a/parquet/src/arrow/buffer/offset_buffer.rs
+++ b/parquet/src/arrow/buffer/offset_buffer.rs
@@ -156,16 +156,19 @@ impl<I: OffsetSizeTrait> OffsetBuffer<I> {
 
     fn build_generic_byte_view(self) -> GenericByteViewBuilder<BinaryViewType> {
         let mut builder = GenericByteViewBuilder::<BinaryViewType>::with_capacity(self.len());
-        let mut values = self.values;
+        let buffer = self.values.into();
+        builder.append_block(buffer);
         for window in self.offsets.windows(2) {
             let start = window[0];
             let end = window[1];
             let len = (end - start).to_usize().unwrap();
-            let b = values.drain(..len).collect::<Vec<u8>>();
-            if b.is_empty() {
-                builder.append_null();
+
+            if len != 0 {
+                builder
+                    .try_append_view(0, start.as_usize() as u32, len as u32)
+                    .unwrap();
             } else {
-                builder.append_value(b);
+                builder.append_null();
             }
         }
         builder

From 4669ae4cf24f860acb855ca67bad564399d03847 Mon Sep 17 00:00:00 2001
From: Xiangpeng Hao <haoxiangpeng123@gmail.com>
Date: Thu, 13 Jun 2024 00:01:07 -0400
Subject: [PATCH 2/2] avoid hard code block id

---
 parquet/src/arrow/buffer/offset_buffer.rs | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/parquet/src/arrow/buffer/offset_buffer.rs b/parquet/src/arrow/buffer/offset_buffer.rs
index 89484181d0cc..181e69c669a4 100644
--- a/parquet/src/arrow/buffer/offset_buffer.rs
+++ b/parquet/src/arrow/buffer/offset_buffer.rs
@@ -157,7 +157,7 @@ impl<I: OffsetSizeTrait> OffsetBuffer<I> {
     fn build_generic_byte_view(self) -> GenericByteViewBuilder<BinaryViewType> {
         let mut builder = GenericByteViewBuilder::<BinaryViewType>::with_capacity(self.len());
         let buffer = self.values.into();
-        builder.append_block(buffer);
+        let block = builder.append_block(buffer);
         for window in self.offsets.windows(2) {
             let start = window[0];
             let end = window[1];
@@ -165,7 +165,7 @@ impl<I: OffsetSizeTrait> OffsetBuffer<I> {
 
             if len != 0 {
                 builder
-                    .try_append_view(0, start.as_usize() as u32, len as u32)
+                    .try_append_view(block, start.as_usize() as u32, len as u32)
                     .unwrap();
             } else {
                 builder.append_null();