从Realm单条记录聚合统计数据的性能优化求助

1. 用Realm原生聚合查询替代遍历 #

Realm内置sum、average等聚合函数，直接针对年份分组查询，无需将所有WorkDay对象加载到内存遍历，这是最直接的性能优化手段。

示例代码（Swift）：

// 先获取所有存在的年份，避免重复计算
let years = realm.objects(WorkDay.self)
    .distinct(by: ["startTime.year"])
    .compactMap { $0.startTime.year }

// 遍历年份，直接查询对应年份的聚合数据
var yearlyStats: [TimePeriodData] = []
for year in years {
    let yearPredicate = NSPredicate(format: "startTime.year == %d", year)
    let yearObjects = realm.objects(WorkDay.self).filter(yearPredicate)
    
    let totalIncome = yearObjects.sum(ofProperty: "totalIncome")
    let averageIncome = yearObjects.average(ofProperty: "totalIncome") ?? 0
    let recordCount = yearObjects.count
    
    yearlyStats.append(TimePeriodData(year: year, totalIncome: totalIncome, averageIncome: averageIncome, recordCount: count))
}

注：给WorkDay的startTime字段添加索引（@Persisted(indexed: true)），能大幅提升年份筛选的查询速度

2. 预计算缓存聚合数据 #

如果需要频繁查询年度统计，可在Realm中新增YearSummary对象，每当添加/更新/删除WorkDay时，同步更新对应的YearSummary记录，后续查询直接读取缓存好的统计值，完全避免实时计算。

定义YearSummary模型：

class YearSummary: Object {
    @Persisted(primaryKey: true) var year: Int
    @Persisted var totalIncome: Double
    @Persisted var averageIncome: Double
    @Persisted var recordCount: Int
}

同步更新逻辑示例：

// 添加WorkDay时同步更新YearSummary
func addWorkDay(_ workDay: WorkDay) {
    try! realm.write {
        realm.add(workDay)
        
        let year = workDay.startTime.year
        if let summary = realm.object(ofType: YearSummary.self, forPrimaryKey: year) {
            summary.totalIncome += workDay.totalIncome
            summary.recordCount += 1
            summary.averageIncome = summary.totalIncome / Double(summary.recordCount)
        } else {
            let newSummary = YearSummary()
            newSummary.year = year
            newSummary.totalIncome = workDay.totalIncome
            newSummary.recordCount = 1
            newSummary.averageIncome = workDay.totalIncome
            realm.add(newSummary)
        }
    }
}

后续查询直接读取缓存：

let yearlyStats = realm.objects(YearSummary.self).sorted(byKeyPath: "year")

3. 减少不必要的对象加载 #

如果必须遍历WorkDay对象，确保只加载需要的字段，而非整个对象。用value(forKey:)或投影查询降低内存开销：

let yearPredicate = NSPredicate(format: "startTime.year == %d", targetYear)
// 仅提取totalIncome字段，避免加载完整WorkDay对象
let incomes = realm.objects(WorkDay.self)
    .filter(yearPredicate)
    .value(forKey: "totalIncome") as! [Double]

let totalIncome = incomes.reduce(0, +)
let averageIncome = incomes.isEmpty ? 0 : totalIncome / Double(incomes.count)

内容的提问来源于stack exchange，提问作者Brad