构建数据质量校验器

项目概述

本项目实现一个面向 data quality enforcement 的完整系统，核心要解决的问题是 checking a dataset against rules for nulls, ranges, and uniqueness before it is consumed downstream。整体由 Rule engine、Column profiler、Uniqueness checker、Range checker、Report emitter 五个模块组成，每个模块单一职责、可独立测试，最后通过集成测试验证端到端正确性。设计上把 a uniqueness rule over a column with whitespace variants that treats trimmed and untrimmed values as distinct 作为第一类需求先行处理，而非事后再补。

架构与关键决策

数据模型为 a set of rules evaluated per column, producing a pass-fail report with counts of violating rows and sample offenders。Rule engine 定义其余阶段共同依赖的数据形状并在边界处校验输入；Column profiler 承载 data quality enforcement 的正确性逻辑；Uniqueness checker 专门处理 a uniqueness rule over a column with whitespace variants that treats trimmed and untrimmed values as distinct，在失败路径上给出明确的拒绝、重试或降级策略；Range checker 提供可观测性，每次操作输出带关联 id 的结构化事件；Report emitter 在过载时按既定方式降级，避免过载演变为数据损坏。配置与代码分离，使风险变更可经 diff 审阅。

实施与测试

按五个阶段依次实现 Rule engine、Column profiler、Uniqueness checker、Range checker、Report emitter，每阶段配套聚焦单测。集成测试组合各模块并驱动真实 data quality enforcement 场景：先验证正常路径，再刻意触发 a uniqueness rule over a column with whitespace variants that treats trimmed and untrimmed values as distinct 并断言系统降级而非损坏。所有依赖均在进程内或被桩替换，用例在 CI 中确定可复现。

限制与后续

• 单区域设计，多区域复制超出范围。
• Rule engine 默认进程内，未水平共享。
• 无鉴权层，假定部署在强制身份的边缘之后。

代码示例

# setup for data-quality-validator
pnpm init
pnpm add fastify csv-parse
pnpm add -D typescript tsx vitest

data_quality_validator/
├── src/index.ts
├── src/rule-engine.ts
├── src/column-profiler.ts
├── tests/data_quality_validator.test.ts
└── package.json

// Rule engine: data quality enforcement core
export function ruleengine(input: RuleengineInput): RuleengineResult {
  // Build the rule engine that anchors data quality enforcement and defines the core data shape.
  return processRuleengine(input);
}

// Column profiler — handles data quality enforcement Add the column profiler so
export class Columnprofiler {
  constructor(private opts: ColumnprofilerOptions) {}
  async run(ctx: Context): Promise<ColumnprofilerOutput> {
    return this.execute(ctx);
  }
}

// Uniqueness checker: data quality enforcement stage
export const uniquenesschecker = (next: Next) => async (req: Request) => {
  const out = await transform(req);
  return next(out);
};

// Range checker for data quality enforcement
export async function rangechecker(stream: AsyncIterable<Chunk>): Promise<Aggregate> {
  let acc = init();
  for await (const c of stream) acc = merge(acc, c);
  return acc;
}

// Report emitter: data quality enforcement finalization
export function reportemitter(state: State): Result {
  validate(state);
  return finalize(state);
}

# run the data quality enforcement test suite
pnpm vitest run