使用reduce处理数组 实现指定Description项不聚合单独输出

原有实现 #

当前通过reduce实现数组字段提取与聚合，代码如下：

data_proces = Object.values(selectedData.reduce((r, { Code, Description, Price}) => {    
    r[Description] ??= { Code, Description, Units: 0, Price: 0 , Total: 0};
    r[Description].Code= Code;
    r[Description].Units++;
    r[Description].Price = Price ;
    r[Description].Total += Price ;
    return r;
}, {}));

现有输出 #

上述代码对所有条目按Description做全量聚合，输出结果如下：

[
    {
        "Code": 0,
        "Description": "No Factured Act",
        "Units": 2,
        "Price": 0,
        "Total": 0
    },
    {
        "Code": 1,
        "Description": "Autopsy",
        "Units": 1,
        "Price": 44,
        "Total": 44
    },
    {
        "Code": 2,
        "Description": "Simple Biopsy",
        "Units": 3,
        "Price": 29,
        "Total": 87
    },
    {
        "Code": 1,
        "Description": "Citology",
        "Units": 4,
        "Price": 15,
        "Total": 60
    },
    {
        "Code": " -",
        "Description": "Free Act",
        "Units": 2,
        "Price": 56789,
        "Total": 91356
    }
]

需求规则 #

需要调整聚合逻辑，两类条目执行不同处理规则：

• Description为No Factured Act、Free Act的条目：不做聚合，每条原始记录单独生成独立对象，单条对象Units固定为1，Total等于当前条目的Price
• 其余Description的条目：保持原有聚合逻辑，按Description分组，累加Units、Total字段，Code、Price取分组内最新遍历到的值

期望输出结构参考（修正原示例的JSON语法笔误）：

[
    {
        "Code": 0,
        "Description": "No Factured Act",
        "Units": 1,
        "Price": 0,
        "Total": 0
    },
    {
        "Code": 0,
        "Description": "No Factured Act",
        "Units": 1,
        "Price": 0,
        "Total": 0
    },
    {
        "Code": 1,
        "Description": "Autopsy",
        "Units": 1,
        "Price": 44,
        "Total": 44
    },
    {
        "Code": 2,
        "Description": "Simple Biopsy",
        "Units": 3,
        "Price": 29,
        "Total": 87
    },
    {
        "Code": 1,
        "Description": "Citology",
        "Units": 4,
        "Price": 15,
        "Total": 60
    },
    {
        "Code": " -",
        "Description": "Free Act",
        "Units": 1,
        "Price": 34567,
        "Total": 34567
    },
    {
        "Code": " -",
        "Description": "Free Act",
        "Units": 1,
        "Price": 56789,
        "Total": 56789
    }
]

实现代码 #

核心调整点是把reduce的累加器从空对象改为空数组，遍历过程中加分支判断区分两类条目即可，完整实现如下：

// 定义不需要聚合的Description集合，判断时间复杂度O(1)
const skipAggregateDesc = new Set(["No Factured Act", "Free Act"]);

const data_proces = selectedData.reduce((result, { Code, Description, Price }) => {
    // 非聚合条目直接生成独立对象推入结果
    if (skipAggregateDesc.has(Description)) {
        result.push({
            Code,
            Description,
            Units: 1,
            Price,
            Total: Price
        });
        return result;
    }

    // 聚合条目先查找是否已存在同Description的分组
    const existingGroup = result.find(item => item.Description === Description);
    if (existingGroup) {
        // 已存在分组则更新字段、做累加
        existingGroup.Code = Code;
        existingGroup.Units++;
        existingGroup.Price = Price;
        existingGroup.Total += Price;
    } else {
        // 不存在分组则新建初始条目
        result.push({
            Code,
            Description,
            Units: 1,
            Price,
            Total: Price
        });
    }
    return result;
}, []);

性能提示：如果待处理数据量超过1000条，可以额外维护一个Description对应数组索引的映射对象，替代find遍历查找，将分组查找的复杂度从O(n)降到O(1)；中小数据量场景下上述写法可读性更高，没有性能问题。

内容的提问来源于stack exchange，提问作者Javier