shardingsphere-proxy: 5.1.1 support interpolateParams=false, but shardingsphere-proxy: 5.5.2 do not support.

[penghuazhou]

Feature Request

shardingsphere-proxy: 5.1.1 support interpolateParams=false, but shardingsphere-proxy: 5.5.2 do not support.

Describe the feature you would like.

shardingsphere-proxy: 5.5.2

if use "sharding:sharding@tcp(127.0.0.1:23307)/sharding_db?charset=utf8mb4&parseTime=True&loc=Local&interpolateParams=false", it will throw error below, if "interpolateParams=true", it will success.

zego@192 claude-flow-demo % go run sharding.go

2026/02/09 16:42:31 /Users/zego/workspace/claude-flow-demo/sharding.go:29 Error 20024 (44000): Sharding value '[B@78a67186' must implements Comparable.
[12.818ms] [rows:-] SELECT deleted_at FROM tbl_group WHERE app_id = 0 AND group_id = '0' AND deleted_at > 0 LIMIT 1
2026/02/09 16:42:31 Query failed: Error 20024 (44000): Sharding value '[B@78a67186' must implements Comparable.

go client

package main

import (
	"fmt"
	"log"

	"gorm.io/driver/mysql"
	"gorm.io/gorm"
	"gorm.io/gorm/logger"
)

func main() {
	// Database connection
	dsn := "sharding:sharding@tcp(192.168.12.213:23307)/sharding_db?charset=utf8mb4&parseTime=True&loc=Local&interpolateParams=false"

	db, err := gorm.Open(mysql.Open(dsn), &gorm.Config{
		Logger: logger.Default.LogMode(logger.Info),
	})
	if err != nil {
		log.Fatalf("Failed to connect: %v", err)
	}

	// Simple query
	var result []struct {
		DeletedAt int64 `gorm:"column:deleted_at"`
	}

	rawSQL := "SELECT deleted_at FROM tbl_group WHERE app_id = ? AND group_id = ? AND deleted_at > 0 LIMIT 1"
	err = db.Raw(rawSQL, int64(0), "0").Scan(&result).Error

	if err != nil {
		log.Fatalf("Query failed: %v", err)
	}

	fmt.Printf("✅ Query succeeded, results: %+v\n", result)
}

database-sharding.yaml

databaseName: sharding_db

dataSources:
  ds_0:
    url: jdbc:mysql://127.0.0.1:3306/sharding?serverTimezone=UTC&useSSL=false&charset=utf8mb4&parseTime=True&allowPublicKeyRetrieval=true
    username: sharding
    password: sharding
    connectionTimeoutMilliseconds: 30000
    idleTimeoutMilliseconds: 60000
    maxLifetimeMilliseconds: 1800000
    maxPoolSize: 500
    minPoolSize: 10
    customPoolProps:
      leakDetectionThreshold: 30000
      registerMbeans: true

rules:
- !SHARDING
  tables:
    tbl_group:
      actualDataNodes: ds_0.tbl_group_${0}_${0}
      tableStrategy:
        complex:
          shardingColumns: app_id,group_id
          shardingAlgorithmName: tbl_group_strategy
  defaultDatabaseStrategy:
    none:
  defaultTableStrategy:
    none:
  defaultKeyGenerateStrategy:
    keyGeneratorName: snowflake
    column: id

  shardingAlgorithms:
    tbl_group_strategy:
      type: COMPLEX_INLINE
      props:
        algorithm-expression: tbl_group_${app_id}_${Math.abs(group_id.hashCode()) % 2}

  keyGenerators:
    snowflake:
      type: SNOWFLAKE
      props:
        worker-id: 561

create table

CREATE TABLE `tbl_group` (
  `app_id` bigint NOT NULL COMMENT '应用ID',
  `group_id` varchar(128) COLLATE utf8mb4_unicode_ci NOT NULL COMMENT '群组ID',
  `deleted_at` bigint NOT NULL DEFAULT '0' COMMENT '删除时间(毫秒时间戳,0表示未删除)'
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci;

[tiya-9975]

Thanks for the detailed reproduction steps and example @penghuazhou 👍

From the error message (Sharding value '[B@…' must implements Comparable), this looks like the sharding value is being treated as a byte[] when interpolateParams=false (prepared statement mode), which would explain why the COMPLEX_INLINE algorithm fails in 5.5.2 but worked in 5.1.1.

A few clarifying questions that may help triage this as a regression:

Does this reproduce with ShardingSphere-JDBC 5.5.2, or only with Proxy?

Does it fail for both app_id and group_id, or only the VARCHAR sharding column?

If you log parameter types on the proxy side, is group_id being received as byte[] instead of String?

If this behavior worked in 5.1.1 with the same config, it likely indicates a regression in prepared-statement parameter handling for complex sharding algorithms. Happy to help test or narrow this down further.

[penghuazhou]

@tiya-9975
Does it fail for both app_id and group_id, or only the VARCHAR sharding column?
it fail only the VARCHAR sharding column.

If this behavior worked in 5.1.1 with the same config, it likely indicates a regression in prepared-statement parameter handling for complex sharding algorithms. Happy to help test or narrow this down further.
this behavior worked in 5.1.1

[penghuazhou]

@terrymanu @tiya-9975 It also not solve in shardingsphere-proxy: 5.5.3. I throw error below.

2026/03/03 16:39:49 /Users/zego/GolandProjects/tcpDemo/shard/shar.go:29 Error 20024 (44000): Sharding value '[B@418aba6b' must implements Comparable.
[12.734ms] [rows:-] SELECT deleted_at FROM tbl_group WHERE app_id = 0 AND group_id = '0' AND deleted_at > 0 LIMIT 1
2026/03/03 16:39:49 Query failed: Error 20024 (44000): Sharding value '[B@418aba6b' must implements Comparable.

package main

import (
	"fmt"
	"log"

	"gorm.io/driver/mysql"
	"gorm.io/gorm"
	"gorm.io/gorm/logger"
)

func main() {
	// Database connection
	dsn := "sharding:sharding@tcp(192.168.12.213:23307)/sharding_db?charset=utf8mb4&parseTime=True&loc=Local&interpolateParams=false"

	db, err := gorm.Open(mysql.Open(dsn), &gorm.Config{
		Logger: logger.Default.LogMode(logger.Info),
	})
	if err != nil {
		log.Fatalf("Failed to connect: %v", err)
	}

	// Simple query
	var result []struct {
		DeletedAt int64 `gorm:"column:deleted_at"`
	}

	rawSQL := "SELECT deleted_at FROM tbl_group WHERE app_id = ? AND group_id = ? AND deleted_at > 0 LIMIT 1"
	err = db.Raw(rawSQL, int64(0), "0").Scan(&result).Error

	if err != nil {
		log.Fatalf("Query failed: %v", err)
	}

	fmt.Printf("✅ Query succeeded, results: %+v\n", result)
}

[penghuazhou]

@terrymanu, @tiya-9975 I download shardingsphere-proxy:5.5.3 from https://www.apache.org/dyn/closer.lua/shardingsphere/5.5.3/apache-shardingsphere-5.5.3-shardingsphere-proxy-bin.tar.gz.