From 7d96ea6987d6ee7b2d4385fb267f161e06e56e7d Mon Sep 17 00:00:00 2001
From: "xiaolei.zl" <xiaolei.zl@alibaba-inc.com>
Date: Fri, 13 Dec 2024 15:42:38 +0800
Subject: [PATCH] cutoff string data for var_char

Committed-by: xiaolei.zl from Dev container
---
 flex/utils/property/column.cc |  7 +++++--
 flex/utils/property/column.h  | 13 +++++++++----
 flex/utils/property/types.h   |  5 +++--
 3 files changed, 17 insertions(+), 8 deletions(-)

diff --git a/flex/utils/property/column.cc b/flex/utils/property/column.cc
index a2581be4caa7..cb29b9c8c9fc 100644
--- a/flex/utils/property/column.cc
+++ b/flex/utils/property/column.cc
@@ -169,11 +169,14 @@ std::shared_ptr<ColumnBase> CreateColumn(
       return std::make_shared<DayColumn>(strategy);
     } else if (type == PropertyType::kStringMap) {
       return std::make_shared<DefaultStringMapColumn>(strategy);
-    } else if (type == PropertyType::kStringView) {
-      return std::make_shared<StringColumn>(strategy);
     } else if (type.type_enum == impl::PropertyTypeImpl::kVarChar) {
+      // We must check is varchar first, because in implementation of
+      // PropertyType::operator==(const PropertyType& other), we string_view is
+      // equal to varchar.
       return std::make_shared<StringColumn>(
           strategy, type.additional_type_info.max_length);
+    } else if (type == PropertyType::kStringView) {
+      return std::make_shared<StringColumn>(strategy);
     } else if (type.type_enum == impl::PropertyTypeImpl::kRecordView) {
       return std::make_shared<RecordViewColumn>(sub_types);
     } else {
diff --git a/flex/utils/property/column.h b/flex/utils/property/column.h
index a556857e77cc..393ba02f7bd2 100644
--- a/flex/utils/property/column.h
+++ b/flex/utils/property/column.h
@@ -499,12 +499,17 @@ class TypedColumn<std::string_view> : public ColumnBase {
   PropertyType type() const override { return PropertyType::Varchar(width_); }
 
   void set_value(size_t idx, const std::string_view& val) {
+    auto copied_val = val;
+    if (copied_val.size() >= width_) {
+      LOG(WARNING) << "String length exceeds the maximum length: " << width_;
+      copied_val = copied_val.substr(0, width_);
+    }
     if (idx >= basic_size_ && idx < basic_size_ + extra_size_) {
-      size_t offset = pos_.fetch_add(val.size());
-      extra_buffer_.set(idx - basic_size_, offset, val);
+      size_t offset = pos_.fetch_add(copied_val.size());
+      extra_buffer_.set(idx - basic_size_, offset, copied_val);
     } else if (idx < basic_size_) {
-      size_t offset = basic_pos_.fetch_add(val.size());
-      basic_buffer_.set(idx, offset, val);
+      size_t offset = basic_pos_.fetch_add(copied_val.size());
+      basic_buffer_.set(idx, offset, copied_val);
     } else {
       LOG(FATAL) << "Index out of range";
     }
diff --git a/flex/utils/property/types.h b/flex/utils/property/types.h
index ad7eb1f36f5a..ae82c000d210 100644
--- a/flex/utils/property/types.h
+++ b/flex/utils/property/types.h
@@ -1408,9 +1408,10 @@ struct convert<gs::PropertyType> {
           if (config["string"]["var_char"]["max_length"]) {
             property_type = gs::PropertyType::Varchar(
                 config["string"]["var_char"]["max_length"].as<int32_t>());
+          } else {
+            property_type = gs::PropertyType::Varchar(
+                gs::PropertyType::STRING_DEFAULT_MAX_LENGTH);
           }
-          property_type = gs::PropertyType::Varchar(
-              gs::PropertyType::STRING_DEFAULT_MAX_LENGTH);
         } else {
           LOG(ERROR) << "Unrecognized string type";
         }