fix: BulkIsValid non-nullable fall-through in ChunkedColumn/ProxyChunkColumn (#51598)

The `if (!nullable_)` branch had no `return`, so a non-nullable column
fell through into the nullable block and invoked the callback a second
time per row (latent — currently masked by callers that guard with `if
(!IsNullable()) return`). Added the missing `return`.

issue: #51385

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Signed-off-by: xiaofanluan <xf@hjjaq.com>
Co-authored-by: xiaofanluan <xf@hjjaq.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

internal/core/src/mmap/ChunkedColumn.h

@@ -175,6 +175,10 @@ class ChunkedColumnBase : public ChunkedColumnInterface {
                     fn(true, i);
                 }
             }
+            // Non-nullable is fully handled above; without this return the
+            // control falls through into the nullable block and invokes fn a
+            // second time per row.
+            return;
         }
         // nullable:
         if (offsets == nullptr) {

internal/core/src/mmap/ChunkedColumnGroup.h

@@ -269,6 +269,10 @@ class ProxyChunkColumn : public ChunkedColumnInterface {
                     fn(true, i);
                 }
             }
+            // Non-nullable is fully handled above; without this return the
+            // control falls through into the nullable block and invokes fn a
+            // second time per row.
+            return;
         }
         // nullable:
         if (count == 0) {

internal/core/src/mmap/ChunkedColumnGroupTest.cpp

@@ -274,6 +274,29 @@ TEST_F(ChunkedColumnGroupTest, ProxyChunkColumn) {
         &offset,
         1);
 
+    // Regression: a non-nullable column must invoke the callback EXACTLY once
+    // per row. Before the missing-return fix, the !nullable_ branch fell
+    // through into the nullable branch and invoked the callback a second time
+    // per row.
+    {
+        std::vector<int64_t> all_offsets = {0, 1, 2, 3, 4};
+        std::unordered_map<size_t, int> call_count;
+        proxy_int64->BulkIsValid(
+            nullptr,
+            [&](bool is_valid, size_t i) {
+                EXPECT_TRUE(is_valid);
+                call_count[i]++;
+            },
+            all_offsets.data(),
+            static_cast<int64_t>(all_offsets.size()));
+        ASSERT_EQ(call_count.size(), all_offsets.size());
+        for (const auto& kv : call_count) {
+            EXPECT_EQ(kv.second, 1)
+                << "row " << kv.first << " callback invoked " << kv.second
+                << " times";
+        }
+    }
+
     // Test string proxy
     auto proxy_string = std::make_shared<ProxyChunkColumn>(
         column_group, FieldId(2), string_field_meta);