From fdd164c0fa319e4d8520c810c827267406a85ec6 Mon Sep 17 00:00:00 2001
From: "Lei, HUANG" <6406592+v0y4g3r@users.noreply.github.com>
Date: Tue, 3 Jun 2025 16:25:02 +0800
Subject: [PATCH] fix(mito): revert initial builder capacity for
 TimeSeriesMemtable (#6231)

* fix/initial-builder-cap:
 ### Enhance Series Initialization and Capacity Management

 - **`simple_bulk_memtable.rs`**: Updated the `Series` initialization to use `with_capacity` with a specified capacity of 8192, improving memory management.
 - **`time_series.rs`**: Introduced `with_capacity` method in `Series` to allow custom initial capacity for `ValueBuilder`. Adjusted `INITIAL_BUILDER_CAPACITY` to 16 for more efficient memory usage. Added a new `new` method to maintain backward compatibility.

* fix/initial-builder-cap:
 ### Adjust Memory Allocation in Memtable

 - **`simple_bulk_memtable.rs`**: Reduced the initial capacity of `Series` from 8192 to 1024 to optimize memory usage.
 - **`time_series.rs`**: Decreased `INITIAL_BUILDER_CAPACITY` from 16 to 4 to improve efficiency in vector building.
---
 src/mito2/src/memtable/simple_bulk_memtable.rs |  2 +-
 src/mito2/src/memtable/time_series.rs          | 10 +++++++---
 2 files changed, 8 insertions(+), 4 deletions(-)

diff --git a/src/mito2/src/memtable/simple_bulk_memtable.rs b/src/mito2/src/memtable/simple_bulk_memtable.rs
index ac7a4d4364..7ea60df05f 100644
--- a/src/mito2/src/memtable/simple_bulk_memtable.rs
+++ b/src/mito2/src/memtable/simple_bulk_memtable.rs
@@ -65,7 +65,7 @@ impl SimpleBulkMemtable {
         } else {
             dedup
         };
-        let series = RwLock::new(Series::new(&region_metadata));
+        let series = RwLock::new(Series::with_capacity(&region_metadata, 1024));
 
         Self {
             id,
diff --git a/src/mito2/src/memtable/time_series.rs b/src/mito2/src/memtable/time_series.rs
index e9ca61fffb..b6e6db2c64 100644
--- a/src/mito2/src/memtable/time_series.rs
+++ b/src/mito2/src/memtable/time_series.rs
@@ -60,7 +60,7 @@ use crate::region::options::MergeMode;
 use crate::row_converter::{DensePrimaryKeyCodec, PrimaryKeyCodecExt};
 
 /// Initial vector builder capacity.
-const INITIAL_BUILDER_CAPACITY: usize = 1024 * 8;
+const INITIAL_BUILDER_CAPACITY: usize = 4;
 
 /// Vector builder capacity.
 const BUILDER_CAPACITY: usize = 512;
@@ -663,15 +663,19 @@ pub(crate) struct Series {
 }
 
 impl Series {
-    pub(crate) fn new(region_metadata: &RegionMetadataRef) -> Self {
+    pub(crate) fn with_capacity(region_metadata: &RegionMetadataRef, builder_cap: usize) -> Self {
         Self {
             pk_cache: None,
-            active: ValueBuilder::new(region_metadata, INITIAL_BUILDER_CAPACITY),
+            active: ValueBuilder::new(region_metadata, builder_cap),
             frozen: vec![],
             region_metadata: region_metadata.clone(),
         }
     }
 
+    pub(crate) fn new(region_metadata: &RegionMetadataRef) -> Self {
+        Self::with_capacity(region_metadata, INITIAL_BUILDER_CAPACITY)
+    }
+
     pub fn is_empty(&self) -> bool {
         self.active.len() == 0 && self.frozen.is_empty()
     }