ddl: Fix default value filling with finer granularity (#10682) (#10691)

close #10680

ddl: Fix default value filling with finer granularity
- Tightens addDefaultValueToColumnIfPossible
  - For NOT NULL missing columns, force_decode=false now returns false unless there is an origin default.
  - This forces a schema sync instead of silently filling 0.
  - force_decode=true still fills a default value (best-effort).

Signed-off-by: JaySon-Huang <tshent@qq.com>

Co-authored-by: JaySon-Huang <tshent@qq.com>

Author: ti-chi-bot

dbms/src/Storages/KVStore/Decode/RegionBlockReader.cpp

@@ -179,6 +179,7 @@ bool RegionBlockReader::readImpl(Block & block, const ReadList & data_list, bool
 {
     VersionColResolver<ReadList> version_col_resolver;
     version_col_resolver.check(block, schema_snapshot->column_defines->size());
+    // The column_ids to read according to schema_snapshot, each elem is (column_id, block_pos)
     const auto & read_column_ids = schema_snapshot->getColId2BlockPosMap();
     const auto & pk_column_ids = schema_snapshot->pk_column_ids;
     const auto & pk_pos_map = schema_snapshot->pk_pos_map;
@@ -269,6 +270,8 @@ bool RegionBlockReader::readImpl(Block & block, const ReadList & data_list, bool
             else
             {
                 // Parse column value from encoded value
+                // Decode the column_ids from `column_ids_iter` to `read_column_ids.end()`
+                // and insert into `block` at position starting from `next_column_pos`
                 if (!appendRowToBlock(
                         *value_ptr,
                         column_ids_iter,
@@ -299,12 +302,12 @@ bool RegionBlockReader::readImpl(Block & block, const ReadList & data_list, bool
                 if constexpr (pk_type == TMTPKType::INT64)
                     static_cast<ColumnInt64 *>(raw_pk_column)->getData().push_back(handle_value);
                 else if constexpr (pk_type == TMTPKType::UINT64)
-                    static_cast<ColumnUInt64 *>(raw_pk_column)->getData().push_back(UInt64(handle_value));
+                    static_cast<ColumnUInt64 *>(raw_pk_column)->getData().push_back(static_cast<UInt64>(handle_value));
                 else
                 {
                     // The pk_type must be Int32/UInt32 or more narrow type
                     // so cannot tell its' exact type here, just use `insert(Field)`
-                    raw_pk_column->insert(Field(handle_value));
+                    raw_pk_column->insert(Field{handle_value});
                     if (unlikely(raw_pk_column->getInt(index) != handle_value))
                     {
                         if (!force_decode)

dbms/src/Storages/KVStore/tests/gtest_region_block_reader.cpp

@@ -691,5 +691,178 @@ try
 }
 CATCH
 
+TEST_F(RegionBlockReaderTest, ReadFromRegionDefaultValue)
+try
+{
+    // With this table_info, column "c1" is "NOT NULL" and has no origin default
+    TableInfo table_info_c1_not_null_no_origin_default(
+        R"({"cols":[{"id":1,"name":{"L":"c0","O":"c0"},"offset":0,"state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":4,"Tp":1}},{"id":2,"name":{"L":"handle","O":"handle"},"offset":1,"state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":515,"Flen":11,"Tp":3}},{"default":"-56083770","id":7,"name":{"L":"c1","O":"c1"},"offset":2,"state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":1,"Flen":20,"Tp":8}},{"id":4,"name":{"L":"c2","O":"c2"},"offset":3,"origin_default":"0.07954397","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":-1,"Flag":4097,"Flen":12,"Tp":4}},{"id":5,"name":{"L":"c5","O":"c5"},"offset":4,"origin_default":"0","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":10,"Tp":246}},{"default":"247262911","id":6,"name":{"L":"c4","O":"c4"},"offset":5,"origin_default":"247262911","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":10,"Tp":246}}],"id":636,"index_info":[],"is_common_handle":false,"keyspace_id":4294967295,"name":{"L":"t0","O":"t0"},"pk_is_handle":true,"state":5,"tiflash_replica":{"Available":true,"Count":1},"update_timestamp":463845180343844895})",
+        NullspaceID);
+
+    // With this table_info, column "c1" is "NOT NULL" and has no default value
+    TableInfo table_info_c1_not_null_no_default_value(
+        R"({"cols":[{"id":1,"name":{"L":"c0","O":"c0"},"offset":0,"state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":4,"Tp":1}},{"id":2,"name":{"L":"handle","O":"handle"},"offset":1,"state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":515,"Flen":11,"Tp":3}},{"id":7,"name":{"L":"c1","O":"c1"},"offset":2,"state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":4097,"Flen":20,"Tp":8}},{"id":4,"name":{"L":"c2","O":"c2"},"offset":3,"origin_default":"0.07954397","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":-1,"Flag":4097,"Flen":12,"Tp":4}},{"id":5,"name":{"L":"c5","O":"c5"},"offset":4,"origin_default":"0","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":10,"Tp":246}},{"default":"247262911","id":6,"name":{"L":"c4","O":"c4"},"offset":5,"origin_default":"247262911","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":10,"Tp":246}}],"id":636,"index_info":[],"is_common_handle":false,"keyspace_id":4294967295,"name":{"L":"t0","O":"t0"},"pk_is_handle":true,"state":5,"tiflash_replica":{"Available":true,"Count":1},"update_timestamp":463845180343844895})",
+        NullspaceID);
+
+    // With this table_info, column "c1" has the "NOT NULL" flag and has origin default "-56083770"
+    TableInfo table_info_c1_not_null_with_origin_default(
+        R"({"cols":[{"id":1,"name":{"L":"c0","O":"c0"},"offset":0,"state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":4,"Tp":1}},{"id":2,"name":{"L":"handle","O":"handle"},"offset":1,"state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":515,"Flen":11,"Tp":3}},{"origin_default":"-56083770","id":7,"name":{"L":"c1","O":"c1"},"offset":2,"state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":1,"Flen":20,"Tp":8}},{"id":4,"name":{"L":"c2","O":"c2"},"offset":3,"origin_default":"0.07954397","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":-1,"Flag":4097,"Flen":12,"Tp":4}},{"id":5,"name":{"L":"c5","O":"c5"},"offset":4,"origin_default":"0","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":10,"Tp":246}},{"default":"247262911","id":6,"name":{"L":"c4","O":"c4"},"offset":5,"origin_default":"247262911","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":10,"Tp":246}}],"id":636,"index_info":[],"is_common_handle":false,"keyspace_id":4294967295,"name":{"L":"t0","O":"t0"},"pk_is_handle":true,"state":5,"tiflash_replica":{"Available":true,"Count":1},"update_timestamp":463845180343844895})",
+        NullspaceID);
+
+    // With this table_info, column "c1" has the "NOT NULL" flag and has origin default "-56083770", but the state is not public
+    TableInfo table_info_c1_not_null_with_origin_default_non_public(
+        R"({"cols":[{"id":1,"name":{"L":"c0","O":"c0"},"offset":0,"state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":4,"Tp":1}},{"id":2,"name":{"L":"handle","O":"handle"},"offset":1,"state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":515,"Flen":11,"Tp":3}},{"origin_default":"-56083770","id":7,"name":{"L":"c1","O":"c1"},"offset":2,"state":3,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":1,"Flen":20,"Tp":8}},{"id":4,"name":{"L":"c2","O":"c2"},"offset":3,"origin_default":"0.07954397","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":-1,"Flag":4097,"Flen":12,"Tp":4}},{"id":5,"name":{"L":"c5","O":"c5"},"offset":4,"origin_default":"0","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":10,"Tp":246}},{"default":"247262911","id":6,"name":{"L":"c4","O":"c4"},"offset":5,"origin_default":"247262911","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":10,"Tp":246}}],"id":636,"index_info":[],"is_common_handle":false,"keyspace_id":4294967295,"name":{"L":"t0","O":"t0"},"pk_is_handle":true,"state":5,"tiflash_replica":{"Available":true,"Count":1},"update_timestamp":463845180343844895})",
+        NullspaceID);
+
+    // With this table_info, column "c1" does not have the "NOT NULL" flag
+    TableInfo table_info_c1_nullable(
+        R"({"cols":[{"id":1,"name":{"L":"c0","O":"c0"},"offset":0,"state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":4,"Tp":1}},{"id":2,"name":{"L":"handle","O":"handle"},"offset":1,"state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":515,"Flen":11,"Tp":3}},{"id":7,"name":{"L":"c1","O":"c1"},"offset":2,"state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":20,"Tp":8}},{"id":4,"name":{"L":"c2","O":"c2"},"offset":3,"origin_default":"0.07954397","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":-1,"Flag":4097,"Flen":12,"Tp":4}},{"id":5,"name":{"L":"c5","O":"c5"},"offset":4,"origin_default":"0","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":10,"Tp":246}},{"default":"247262911","id":6,"name":{"L":"c4","O":"c4"},"offset":5,"origin_default":"247262911","state":5,"type":{"Charset":"binary","Collate":"binary","Decimal":0,"Flag":0,"Flen":10,"Tp":246}}],"id":636,"index_info":[],"is_common_handle":false,"keyspace_id":4294967295,"name":{"L":"t0","O":"t0"},"pk_is_handle":true,"state":5,"tiflash_replica":{"Available":true,"Count":1},"update_timestamp":463844343842340870})",
+        NullspaceID);
+
+    RegionID region_id = 4;
+    // the start_key and end_key for table_id = 68
+    String region_start_key(bytesFromHexString("7480000000000002FF7C5F720000000000FA"));
+    String region_end_key(bytesFromHexString("7480000000000002FF7D00000000000000F8"));
+    auto region = makeRegion(region_id, region_start_key, region_end_key);
+    // the hex kv dump from RaftLog
+    std::vector<std::tuple<std::string_view, std::string_view>> kvs = {
+        {
+            "7480000000000002FFA95F728000000000FF0000010000000000FAF9901806DEF7FFDA",
+            "50A380A08892FFF9B706762C8000040000000405060708000F0016001A00BFDC4011A00000000A0080000000000A008000003CA339"
+            "1ABC85",
+        },
+        {
+            "7480000000000002FFA95F728000000000FF0000010000000000FAF9901806DEF7FFD8",
+            "50A680A08892FFF9B706762C8000040000000405060708000F0016001A00BFDC4011A00000000A008033E04D600A008000003CA339"
+            "1ABC85",
+        },
+        {
+            "7480000000000002FFA95F728000000000FF0000020000000000FAF9901806DE33FFE8",
+            "509680B08E92FFF9B706762580000300000004050608000F001600BF720CDD400000000A0080000000010A00800000393C",
+        },
+    };
+    for (const auto & [k, v] : kvs)
+    {
+        region->insert("write", TiKVKey(bytesFromHexString(k)), TiKVValue(bytesFromHexString(v)));
+    }
+
+    auto data_list_read = ReadRegionCommitCache(region, true);
+    ASSERT_TRUE(data_list_read.has_value());
+
+    // Test with `table_info_c1_not_null_no_origin_default`
+    auto decoding_schema = getDecodingStorageSchemaSnapshot(table_info_c1_not_null_no_origin_default);
+    {
+        // force_decode=false can not decode because there are
+        // missing value for column with not null flag.
+        auto reader = RegionBlockReader(decoding_schema);
+        Block res_block = createBlockSortByColumnID(decoding_schema);
+        ASSERT_FALSE(reader.read(res_block, *data_list_read, false));
+    }
+    {
+        // force_decode=true can decode the block, and filling the "zero value" for c1
+        auto reader = RegionBlockReader(decoding_schema);
+        Block res_block = createBlockSortByColumnID(decoding_schema);
+        ASSERT_TRUE(reader.read(res_block, *data_list_read, true));
+        LOG_INFO(
+            Logger::get(),
+            "Decoded block:\n{}",
+            DB::tests::getColumnsContent(res_block.getColumnsWithTypeAndName()));
+        ASSERT_EQ(res_block.getByName("c1").type->getName(), "Int64");
+        // verify the default value is filled correctly
+        ASSERT_COLUMN_EQ( //
+            res_block.getByName("c1"),
+            createColumn<Int64>({-2051270087, -2051270087, 0}));
+    }
+
+    // Test with `table_info_c1_not_null_no_default_value`
+    decoding_schema = getDecodingStorageSchemaSnapshot(table_info_c1_not_null_no_default_value);
+    {
+        // force_decode=false can not decode because there are
+        // missing value for column with not null flag.
+        auto reader = RegionBlockReader(decoding_schema);
+        Block res_block = createBlockSortByColumnID(decoding_schema);
+        ASSERT_FALSE(reader.read(res_block, *data_list_read, false));
+    }
+    {
+        // force_decode=true can decode the block, and filling the "zero value" for c1
+        auto reader = RegionBlockReader(decoding_schema);
+        Block res_block = createBlockSortByColumnID(decoding_schema);
+        ASSERT_TRUE(reader.read(res_block, *data_list_read, true));
+        LOG_INFO(
+            Logger::get(),
+            "Decoded block:\n{}",
+            DB::tests::getColumnsContent(res_block.getColumnsWithTypeAndName()));
+        ASSERT_EQ(res_block.getByName("c1").type->getName(), "Int64");
+        // verify the default value is filled correctly
+        ASSERT_COLUMN_EQ( //
+            res_block.getByName("c1"),
+            createColumn<Int64>({-2051270087, -2051270087, 0}));
+    }
+
+    // Test with `table_info_c1_not_null_with_origin_default`
+    decoding_schema = getDecodingStorageSchemaSnapshot(table_info_c1_not_null_with_origin_default);
+    {
+        // force_decode=false can decode because origin_default exists and NoDefaultValue flag is not set
+        // so RegionBlockReader can use origin_default to fill the missing value`
+        auto reader = RegionBlockReader(decoding_schema);
+        Block res_block = createBlockSortByColumnID(decoding_schema);
+        ASSERT_TRUE(reader.read(res_block, *data_list_read, false));
+        LOG_INFO(
+            Logger::get(),
+            "Decoded block:\n{}",
+            DB::tests::getColumnsContent(res_block.getColumnsWithTypeAndName()));
+        ASSERT_EQ(res_block.getByName("c1").type->getName(), "Int64");
+        // verify the default value is filled correctly
+        ASSERT_COLUMN_EQ( //
+            res_block.getByName("c1"),
+            // the third elem is filled with origin_default
+            createColumn<Int64>({-2051270087, -2051270087, -56083770}));
+    }
+
+    // Test with `table_info_c1_not_null_with_origin_default_non_public`
+    decoding_schema = getDecodingStorageSchemaSnapshot(table_info_c1_not_null_with_origin_default_non_public);
+    {
+        // force_decode=false can not decode because there are
+        // missing value for column with not null flag and the state is not public
+        auto reader = RegionBlockReader(decoding_schema);
+        Block res_block = createBlockSortByColumnID(decoding_schema);
+        ASSERT_FALSE(reader.read(res_block, *data_list_read, false));
+    }
+    {
+        // force_decode=true can decode the block, and filling the default value for c1
+        auto reader = RegionBlockReader(decoding_schema);
+        Block res_block = createBlockSortByColumnID(decoding_schema);
+        ASSERT_TRUE(reader.read(res_block, *data_list_read, true));
+        LOG_INFO(
+            Logger::get(),
+            "Decoded block:\n{}",
+            DB::tests::getColumnsContent(res_block.getColumnsWithTypeAndName()));
+        ASSERT_EQ(res_block.getByName("c1").type->getName(), "Int64");
+        // verify the default value is filled correctly
+        ASSERT_COLUMN_EQ( //
+            res_block.getByName("c1"),
+            // the third elem is filled with origin_default
+            createColumn<Int64>({-2051270087, -2051270087, -56083770}));
+    }
+
+    // Test with `table_info_c1_nullable`
+    decoding_schema = getDecodingStorageSchemaSnapshot(table_info_c1_nullable);
+    {
+        // force_decode=false should be able to decode because c1 is nullable
+        auto reader = RegionBlockReader(decoding_schema);
+        Block res_block = createBlockSortByColumnID(decoding_schema);
+        ASSERT_TRUE(reader.read(res_block, *data_list_read, false));
+        LOG_INFO(
+            Logger::get(),
+            "Decoded block:\n{}",
+            DB::tests::getColumnsContent(res_block.getColumnsWithTypeAndName()));
+        ASSERT_EQ(res_block.getByName("c1").type->getName(), "Nullable(Int64)");
+        // verify the default value is filled with NULL correctly at the last row
+        ASSERT_COLUMN_EQ( //
+            res_block.getByName("c1"),
+            createNullableColumn<Int64>({-2051270087, -2051270087, 0}, {0, 0, 1}));
+    }
+}
+CATCH
 
 } // namespace DB::tests