系统重构时优化关于枚举和字符串

最近在重构同事写的代码,代码设计的非常好,由于涉及到的类比较多,我这里就以一个枚举类来记录关于枚举和字符串的优化(不涉及到业务部分)

代码如下：

package com.vcg.community.data.constant;
public enum InvokerTypeEnum {
    //读操作
    ReadEntityById {
        public String cmdClassName() {return "com.vcg.community.data.command.read.EntityByIdCmd";}
    },
    ReadEntityByIndex {
        public String cmdClassName() {return "com.vcg.community.data.command.read.EntityByIndexCmd";}
    },
    ReadEntityByLink{
        public String cmdClassName() {return "com.vcg.community.data.command.read.EntityByLinkCmd";}
    },
    ReadEntityByColumn{
        public String cmdClassName() {return "com.vcg.community.data.command.read.EntityByColumnCmd";}
    },
    ReadCount {
        public String cmdClassName() {return "com.vcg.community.data.command.read.CountCmd";}
    },
    ReadCountByLink {
        public String cmdClassName() {return "com.vcg.community.data.command.read.CntByLinkCmd";}
    },
    ReadStateByLink {
        public String cmdClassName() {return "com.vcg.community.data.command.read.StateByLinkCmd";}
    },
    Search {
        public String cmdClassName() {return "com.vcg.community.data.command.read.SearchCmd";}
    },
Msg {
    public String cmdClassName() {return "com.vcg.community.data.command.read.MsgCmd";}
},
ReadEntityBySql {
    public String cmdClassName() {return "com.vcg.community.data.command.read.EntityBySqlCmd";}
},
SearchBySql {
    public String cmdClassName() {return "com.vcg.community.data.command.read.SearchBySqlCmd";}
},
QueryCountBySql {
    public String cmdClassName() {return "com.vcg.community.data.command.read.QueryCountBySqlCmd";}
},
QueryIdsBySql {
    public String cmdClassName() {return "com.vcg.community.data.command.read.QueryIdsBySqlCmd";}
},
//写操作
InsertEntitys{
    public String cmdClassName() {return "com.vcg.community.data.command.write.InsertEntitysCmd";}
},
DeleteEntitys{
    public String cmdClassName() {return "com.vcg.community.data.command.write.DeleteEntitysCmd";}
},
DoLinks{
    public String cmdClassName() {return "com.vcg.community.data.command.write.DoLinksCmd";}
},
UndoLinks{
    public String cmdClassName() {return "com.vcg.community.data.command.write.UndoLinksCmd";}
},
WriteOther{
    public String cmdClassName() {return "com.vcg.community.data.command.write.WriteOtherCmd";}
},
ChangeCount{
    public String cmdClassName() {return "com.vcg.community.data.command.write.ChangeCountCmd";}
};
public abstract String cmdClassName() ;

}

上下文代码

InvokerTypeEnum invokeType = inputObject.getInvokeType();
        if (invokeType == null) {
            return new OutputObject(OutputObject.CODE_ERROR, "the attr invoke_type can not be null");
        }
        String className = invokeType.cmdClassName();
        AbstractCommand dp = (AbstractCommand) SpringUtils.getBeanByClassName(className);
        outputObject = dp.checkInputAttr(inputObject);
        if (outputObject.getCode() == OutputObject.CODE_ERROR) {
            return outputObject;
        }
        outputObject = dp.execute(inputObject);
        return outputObject;
    }

public abstract class AbstractCommand {
    @Autowired
    protected RedisStrategy redisStrategy;
    @Autowired
    protected MysqlStrategy mysqlStrategy;
    @Autowired
    protected SearchStrategy searchStrategy;
    public abstract OutputObject execute(InputObject inputObject);
    public abstract OutputObject checkInputAttr(InputObject inputObject);
}

优化后的代码：

public enum InvokerType {
    //读操作
    ReadEntityById,
    ReadEntityByIndex,
    ReadEntityByLink,
    ReadEntityByColumn,
    ReadCount,
    ReadCountByLink,
    ReadStateByLink,
    Search,
    Msg,
    ReadEntityBySql,
    SearchBySql,
    QueryCountBySql,
    QueryIdsBySql,
    //写操作
    InsertEntitys,
    DeleteEntity,
    DoLinks,
    UndoLinks,
    WriteOther,
    ChangeCount;
    String cmdClassName(){return this.getClass().getPackage().getName()+".command."+this.name()+"Cmd";}
}

这样一来,是不是更合理了呢？我们来分析一下

1.首先从代码的数量来说,是不是更精简了一些。
2.形式来说,避开了抽象枚举,是不是更接近普通的枚举设计。
3.性能来说,同事的方式会让JVM产生更多的类似“com.vcg.community.data.command.read/write.xxxxx”的字符串常量(不是new出来的字符串都会在编译期间存入常量池中)
4.系统扩展性,当系统可能会新增该枚举类型时,优化后的方式更利于系统扩展,仅需新增一个枚举值即可。
5.系统稳定性，减少可能由于字符输入(大小写,非复制手敲等原因)错误而造成spring无法加载模板实例的可能性


最终,同事在我的说服下同意了我的修改。