org.springframework.dao.InvalidDataAccessApiUsageException: ERR value is not an integer or out of ra

背景

使用spring redis的increment方法时，爆出如上错误，详细异常栈如下

org.springframework.dao.InvalidDataAccessApiUsageException: ERR value is not an integer or out of range; nested exception is redis.clients.jedis.exceptions.JedisDataException: ERR value is not an integer or out of range
at org.springframework.data.redis.connection.jedis.JedisExceptionConverter.convert(JedisExceptionConverter.java:64)
at org.springframework.data.redis.connection.jedis.JedisExceptionConverter.convert(JedisExceptionConverter.java:41)
at org.springframework.data.redis.PassThroughExceptionTranslationStrategy.translate(PassThroughExceptionTranslationStrategy.java:37)
at org.springframework.data.redis.FallbackExceptionTranslationStrategy.translate(FallbackExceptionTranslationStrategy.java:37)
at org.springframework.data.redis.connection.jedis.JedisConnection.convertJedisAccessException(JedisConnection.java:212)
at org.springframework.data.redis.connection.jedis.JedisConnection.incrBy(JedisConnection.java:1439)
at org.springframework.data.redis.core.DefaultValueOperations$3.doInRedis(DefaultValueOperations.java:66)
at org.springframework.data.redis.core.DefaultValueOperations$3.doInRedis(DefaultValueOperations.java:63)
at org.springframework.data.redis.core.RedisTemplate.execute(RedisTemplate.java:202)
at org.springframework.data.redis.core.RedisTemplate.execute(RedisTemplate.java:164)
at org.springframework.data.redis.core.AbstractOperations.execute(AbstractOperations.java:88)
at org.springframework.data.redis.core.DefaultValueOperations.increment(DefaultValueOperations.java:63)

分析

redis对任何不合法的值，都称为ERR。

也就是说，上述异常的意思是，增加的值（ERR）不是integer或者超出了integer范围，所以才报的错误。
根据同事的建议，只有使用StringRedisSerializer序列化器才能使用incrment方法！那么为什么呢，为什么使用了StringRedisSerializer就可以而使用默认序列化器就不行呢？

实验

我们都知道序列化器是，将key或者value序列化成流，存储到redis。首先看一下RedisSerializer接口

它有两个方法，存储到redis时，使用的是serialize(T t)方法
为了测试各个序列化器的区别，我做了个测试

@Test
    public void testSerializer(){

        RedisTemplate<String, String> template = new RedisTemplate<String, String>();
        template.setConnectionFactory(connectFactory);
        template.afterPropertiesSet();

        ValueOperations<String, String>  operations = template.opsForValue();
        template.setKeySerializer(new StringRedisSerializer());
        template.setValueSerializer(new StringRedisSerializer());
        operations.set("StringRedisSerializer", "1");

        template.setKeySerializer(new GenericToStringSerializer<String>(String.class));
        template.setValueSerializer(new GenericToStringSerializer<String>(String.class));
        operations.set("GenericToStringSerializer", "1");


        template.setKeySerializer(new GenericJackson2JsonRedisSerializer());
        template.setValueSerializer(new GenericJackson2JsonRedisSerializer());
        operations.set("GenericJackson2JsonRedisSerializer", "1");

        template.setKeySerializer(new Jackson2JsonRedisSerializer<String>(String.class));
        template.setValueSerializer(new Jackson2JsonRedisSerializer<String>(String.class));
        operations.set("Jackson2JsonRedisSerializer", "1");

        template.setKeySerializer(new JdkSerializationRedisSerializer());
        template.setValueSerializer(new JdkSerializationRedisSerializer());
        operations.set("JdkSerializationRedisSerializer", "1");


    }

结果

领悟

至此真相大白，

• GenericJackson2JsonRedisSerializer、Jackson2JsonRedisSerializer是先将对象转为json，然后再保存到redis，所以，1在redis中是字符串1，所以无法进行加1
• JdkSerializationRedisSerializer使用的jdk对象序列化，序列化后的值有类信息、版本号等，所以是一个包含很多字母的字符串，所以根本无法加1,这个序列化器跟memcache的序列化规则很像memcache怎样存储的对象
• GenericToStringSerializer、StringRedisSerializer将字符串的值直接转为字节数组，所以保存到redis中是数字，所以可以进行加1

最后，看一下StringRedisSerializer的源代码

GenericToStringSerializer serialize方法

public byte[] serialize(T object) {
        if (object == null) {
            return null;
        }
        String string = converter.convert(object, String.class);
        return string.getBytes(charset);
    }

结论

使用GenericToStringSerializer、StringRedisSerializer序列化器，都可以使用increment方法