[refactor](be) Move adaptive block size predictor to core

be/src/storage/segment/adaptive_block_size_predictor.cpp → be/src/core/block/adaptive_block_size_predictor.cpp

@@ -15,7 +15,7 @@
 // specific language governing permissions and limitations
 // under the License.
 
-#include "storage/segment/adaptive_block_size_predictor.h"
+#include "core/block/adaptive_block_size_predictor.h"
 
 #include <algorithm>
 #include <cstddef>
@@ -46,7 +46,7 @@ void AdaptiveBlockSizePredictor::update(const Block& block) {
     }
 }
 
-size_t AdaptiveBlockSizePredictor::predict_next_rows() {
+size_t AdaptiveBlockSizePredictor::predict_next_rows() const {
     if (_block_size_bytes == 0) {
         return _block_size_rows;
     }

be/src/storage/segment/adaptive_block_size_predictor.h → be/src/core/block/adaptive_block_size_predictor.h

@@ -17,12 +17,7 @@
 
 #pragma once
 
-#include <stddef.h>
-#include <stdint.h>
-
-#include <vector>
-
-#include "storage/olap_common.h"
+#include <cstddef>
 
 namespace doris {
 
@@ -41,12 +36,6 @@ class AdaptiveBlockSizePredictor {
     static constexpr size_t kDefaultProbeRows = 4096;
     static constexpr size_t kDefaultBlockSizeRows = 65535;
 
-    // Per-column metadata for computing segment-level hints.
-    struct ColumnMetadata {
-        ColumnId column_id;
-        uint64_t raw_bytes; // total raw data bytes for this column in the segment
-    };
-
     // |preferred_block_size_bytes|: target total bytes of each output block chunk.
     // |metadata_hint_bytes_per_row|: pre-computed conservative estimate from metadata (e.g.
     //     segment footer or file statistics). 0.0 means no hint available.
@@ -65,7 +54,7 @@ class AdaptiveBlockSizePredictor {
     // Never exceeds |block_size_rows|; never returns less than 1.
     // Uses pre-computed metadata hint for first-call estimate when no history exists.
     // Does NOT modify internal state (_has_history is only flipped by update()).
-    size_t predict_next_rows();
+    size_t predict_next_rows() const;
 
     bool has_history() const { return _has_history; }
 
@@ -84,9 +73,8 @@ class AdaptiveBlockSizePredictor {
     // Whether at least one update() has been called (i.e. we have real measured history).
     bool _has_history = false;
 
-    // Cached conservative metadata estimate computed on the first predict_next_rows() call.
-    // Reused on subsequent first-round predictions (before _has_history is set) to avoid
-    // re-traversing the segment footer on every call.
+    // Conservative metadata estimate provided by the caller at construction.
+    // Reused for predictions before _has_history is set.
     double _metadata_hint_bytes_per_row = 0.0;
 
 #ifdef BE_TEST

be/src/exec/scan/file_scanner.h

@@ -30,6 +30,7 @@
 #include "common/factory_creator.h"
 #include "common/global_types.h"
 #include "common/status.h"
+#include "core/block/adaptive_block_size_predictor.h"
 #include "core/block/block.h"
 #include "exec/operator/file_scan_operator.h"
 #include "exec/scan/file_scan_io_context.h"
@@ -43,7 +44,6 @@
 #include "runtime/runtime_profile.h"
 #include "storage/olap_common.h"
 #include "storage/olap_scan_common.h"
-#include "storage/segment/adaptive_block_size_predictor.h"
 #include "storage/segment/condition_cache.h"
 
 namespace doris {

be/src/exec/scan/file_scanner_v2.h

@@ -26,6 +26,7 @@
 
 #include "common/factory_creator.h"
 #include "common/status.h"
+#include "core/block/adaptive_block_size_predictor.h"
 #include "core/block/block.h"
 #include "exec/operator/file_scan_operator.h"
 #include "exec/scan/scanner.h"
@@ -37,7 +38,6 @@
 #include "gen_cpp/PlanNodes_types.h"
 #include "io/io_common.h"
 #include "runtime/runtime_profile.h"
-#include "storage/segment/adaptive_block_size_predictor.h"
 
 namespace doris {
 

be/src/storage/segment/segment_iterator.cpp

@@ -382,7 +382,6 @@ std::unique_ptr<AdaptiveBlockSizePredictor> SegmentIterator::_make_block_size_pr
     // Collect per-column raw byte metadata from the segment footer for the columns
     // this iterator will actually output (defined by _schema, which is built from
     // _opts.return_columns).
-    std::vector<AdaptiveBlockSizePredictor::ColumnMetadata> col_metadata;
     uint32_t seg_rows = _segment->num_rows();
     uint64_t total_raw_bytes = 0;
     double metadata_hint_bytes_per_row = 0.0;

be/src/storage/segment/segment_iterator.h

@@ -32,6 +32,7 @@
 #include <vector>
 
 #include "common/status.h"
+#include "core/block/adaptive_block_size_predictor.h"
 #include "core/block/block.h"
 #include "core/block/column_with_type_and_name.h"
 #include "core/block/columns_with_type_and_name.h"
@@ -52,7 +53,6 @@
 #include "storage/predicate/column_predicate.h"
 #include "storage/row_cursor.h"
 #include "storage/schema.h"
-#include "storage/segment/adaptive_block_size_predictor.h"
 #include "storage/segment/common.h"
 #include "storage/segment/segment.h"
 #include "util/slice.h"

be/test/storage/segment/adaptive_block_size_predictor_test.cpp → be/test/core/block/adaptive_block_size_predictor_test.cpp

@@ -15,21 +15,23 @@
 // specific language governing permissions and limitations
 // under the License.
 
-#include "storage/segment/adaptive_block_size_predictor.h"
+#include "core/block/adaptive_block_size_predictor.h"
 
 #include <gmock/gmock.h>
 #include <gtest/gtest.h>
 
+#include <algorithm>
+#include <cstdint>
 #include <memory>
-#include <vector>
+#include <string>
+#include <utility>
 
 #include "common/config.h"
 #include "core/block/block.h"
 #include "core/column/column_string.h"
 #include "core/column/column_vector.h"
 #include "core/data_type/data_type_number.h"
 #include "core/data_type/data_type_string.h"
-#include "storage/olap_common.h"
 
 namespace doris {
 
@@ -81,7 +83,6 @@ TEST_F(AdaptiveBlockSizePredictorTest, NoHistoryReturnsMaxRows) {
 
     // After one update the first sample is stored directly (no EWMA blending).
     Block blk = make_int32_block(100);
-    std::vector<ColumnId> cols = {0};
     pred.update(blk);
 
     EXPECT_TRUE(pred.has_history_for_test());
@@ -95,8 +96,6 @@ TEST_F(AdaptiveBlockSizePredictorTest, NoHistoryReturnsMaxRows) {
 TEST_F(AdaptiveBlockSizePredictorTest, EwmaConvergence) {
     AdaptiveBlockSizePredictor pred(kBlockBytes, 0.0);
 
-    std::vector<ColumnId> cols = {0};
-
     // Compute expected bytes-per-row from an actual block so the test does not
     // hard-code internal column memory layout assumptions.
     Block probe = make_string_block(100, 100);
@@ -120,7 +119,6 @@ TEST_F(AdaptiveBlockSizePredictorTest, ZeroRowsBlockIgnored) {
 
     // update() with an empty block must be a no-op.
     Block blk = make_int32_block(0);
-    std::vector<ColumnId> cols = {0};
     pred.update(blk);
 
     EXPECT_FALSE(pred.has_history_for_test());
@@ -139,7 +137,6 @@ TEST_F(AdaptiveBlockSizePredictorTest, DisabledWhenBlockSizeIsZero) {
     AdaptiveBlockSizePredictor pred(0, 0.0);
 
     Block blk = make_int32_block(1000);
-    std::vector<ColumnId> cols = {0};
     pred.update(blk);
 
     // update() still records history even when budget == 0.
@@ -162,7 +159,6 @@ TEST_F(AdaptiveBlockSizePredictorTest, PredictReturnsBlockSizeRowsWhenDisabled)
 
     // Even after update, still returns block_size_rows because block_size_bytes == 0.
     Block blk = make_int32_block(100);
-    std::vector<ColumnId> cols = {0};
     pred.update(blk);
     EXPECT_EQ(pred.predict_next_rows(), pred.block_size_rows_for_test());
 }
@@ -191,7 +187,7 @@ TEST_F(AdaptiveBlockSizePredictorTest, PredictNoHistoryMetadataHint) {
 
     size_t result = pred.predict_next_rows();
 
-    size_t expected = static_cast<size_t>(static_cast<double>(kBlockBytes) / hint_bpr);
+    auto expected = static_cast<size_t>(static_cast<double>(kBlockBytes) / hint_bpr);
     // No history: probe_rows clamps the result.
     expected = std::min(expected, pred.probe_rows_for_test());
     EXPECT_EQ(result, expected);
@@ -237,7 +233,7 @@ TEST_F(AdaptiveBlockSizePredictorTest, PredictWithHistoryNoClamping) {
     pred.set_has_history_for_test(true, 100.0);
 
     size_t result = pred.predict_next_rows();
-    EXPECT_EQ(result, 81u);
+    EXPECT_EQ(result, 81U);
 }
 
 // ── Test: predicted > block_size_rows → clamped to block_size_rows ─────────────
@@ -257,7 +253,7 @@ TEST_F(AdaptiveBlockSizePredictorTest, PredictClampedToOne) {
     // bytes_per_row so large that predicted rounds to 0.
     pred.set_has_history_for_test(true, static_cast<double>(kBlockBytes) * 10.0);
 
-    EXPECT_EQ(pred.predict_next_rows(), 1u);
+    EXPECT_EQ(pred.predict_next_rows(), 1U);
 }
 
 // ── Test: metadata hint with multiple columns ───────────────────────────────
@@ -268,7 +264,7 @@ TEST_F(AdaptiveBlockSizePredictorTest, PredictNoHistoryMultiColumnMetadata) {
     AdaptiveBlockSizePredictor pred(kBlockBytes, hint_bpr);
 
     size_t result = pred.predict_next_rows();
-    size_t expected = static_cast<size_t>(static_cast<double>(kBlockBytes) / hint_bpr);
+    auto expected = static_cast<size_t>(static_cast<double>(kBlockBytes) / hint_bpr);
     // No history: probe_rows clamps the result.
     expected = std::min(expected, pred.probe_rows_for_test());
     EXPECT_EQ(result, expected);
@@ -301,14 +297,14 @@ TEST_F(AdaptiveBlockSizePredictorTest, PredictUsesCustomProbeRowsWithHint) {
 TEST_F(AdaptiveBlockSizePredictorTest, PredictProbeRowsZeroFallsBackToOne) {
     AdaptiveBlockSizePredictor pred(kBlockBytes, 0.0, 0);
 
-    EXPECT_EQ(pred.probe_rows_for_test(), 0u);
-    EXPECT_EQ(pred.predict_next_rows(), 1u);
+    EXPECT_EQ(pred.probe_rows_for_test(), 0U);
+    EXPECT_EQ(pred.predict_next_rows(), 1U);
 }
 
 TEST_F(AdaptiveBlockSizePredictorTest, PredictProbeRowsOneWorks) {
     AdaptiveBlockSizePredictor pred(kBlockBytes, 0.0, 1);
 
-    EXPECT_EQ(pred.predict_next_rows(), 1u);
+    EXPECT_EQ(pred.predict_next_rows(), 1U);
 }
 
 // ── batch_size tests ────────────────────────────────────────────────────────
@@ -318,7 +314,7 @@ TEST_F(AdaptiveBlockSizePredictorTest, DefaultBlockSizeRows) {
 
     EXPECT_EQ(pred.block_size_rows_for_test(),
               AdaptiveBlockSizePredictor::default_block_size_rows_for_test());
-    EXPECT_EQ(pred.block_size_rows_for_test(), 65535u);
+    EXPECT_EQ(pred.block_size_rows_for_test(), 65535U);
 }
 
 TEST_F(AdaptiveBlockSizePredictorTest, CustomBlockSizeRows) {
@@ -351,7 +347,7 @@ TEST_F(AdaptiveBlockSizePredictorTest, BlockSizeRowsDoesNotAffectSmallPrediction
 
     // 100 bytes/row → predicted = 8192/100 = 81 < custom_rows.
     pred.set_has_history_for_test(true, 100.0);
-    EXPECT_EQ(pred.predict_next_rows(), 81u);
+    EXPECT_EQ(pred.predict_next_rows(), 81U);
 }
 
 } // namespace doris