阅读中文版本

Build a Data Diff CLI

Data Diff CLI — a complete walkthrough of data comparison, covering snapshot reader, key indexer, row comparator, and testing.

data_diff_cli/
├── src/index.ts
├── src/snapshot-reader.ts
├── src/key-indexer.ts
├── tests/data_diff_cli.test.ts
└── package.json

// Snapshot reader: data comparison core
export function snapshotreader(input: SnapshotreaderInput): SnapshotreaderResult {
  // Build the snapshot reader that anchors data comparison and defines the core data shape.
  return processSnapshotreader(input);
}

// Key indexer — handles data comparison Add the key indexer so
export class Keyindexer {
  constructor(private opts: KeyindexerOptions) {}
  async run(ctx: Context): Promise<KeyindexerOutput> {
    return this.execute(ctx);
  }
}

// Row comparator: data comparison stage
export const rowcomparator = (next: Next) => async (req: Request) => {
  const out = await transform(req);
  return next(out);
};

// Patch emitter for data comparison
export async function patchemitter(stream: AsyncIterable<Chunk>): Promise<Aggregate> {
  let acc = init();
  for await (const c of stream) acc = merge(acc, c);
  return acc;
}

// Summary reporter: data comparison finalization
export function summaryreporter(state: State): Result {
  validate(state);
  return finalize(state);
}

Project overview

This project builds a data comparison system whose reason for existing is computing the rows that differ between two snapshots of a dataset and emitting a patch. The temptation in data comparison is to bolt the interesting part onto a working prototype and discover the hard cases in production; this build inverts that order and treats a floating-point field that differs by rounding noise and produces a huge, meaningless change set as the first thing designed, not the last.

You will implement Snapshot reader, Key indexer, Row comparator, Patch emitter, and Summary reporter, each as a module with one responsibility and a test that proves it in isolation. The data comparison pipeline is then the wiring that composes them, and the integration test is where the composition is shown to survive a floating-point field that differs by rounding noise and produces a huge, meaningless change set.

The walkthrough is written for a developer who has used a data comparison system but has not had to make one honest. It favors explicit data shapes over clever abstractions, and it names the failure each design choice prevents rather than presenting the choice as a style preference.

Each phase ends with something observable — a snapshot reader that returns the right shape, a key indexer that rejects bad input, an integration test that goes green — so you can stop at any point and still have a data comparison system that works for the cases covered so far.

What you will build

The finished data comparison system exposes five capabilities, each with a focused test:

• Snapshot reader
• Key indexer
• Row comparator
• Patch emitter
• Summary reporter

Snapshot reader is the foundation that every later stage assumes; Key indexer is where data comparison correctness lives; Row comparator handles a floating-point field that differs by rounding noise and produces a huge, meaningless change set; Patch emitter makes the system observable; and Summary reporter is the graceful-degradation layer that keeps overload from becoming corruption.

The build keeps the data comparison logic in ordinary modules you can read in one sitting, and pulls in dependencies only for the parts that are genuinely tedious to hand-roll. That keeps the diff between the tutorial and a real codebase small, and keeps the failure modes visible in code rather than hidden behind a framework.

Architecture and data model

The system is organized around Snapshot reader, Key indexer, and Row comparator, with Patch emitter tying them together. Data flows as two keyed row sets compared by primary key, classifying each row as added, removed, or changed with a field-level diff.

Each component is a module with a pure interface where practical: the same input produces the same output, and side effects are pushed to the edges. That discipline is what makes the phases independently testable, because a test for Row comparator does not need to stand up Snapshot reader and Key indexer to exercise the interesting branch.

The data model is deliberately explicit. Two keyed row sets compared by primary key, classifying each row as added, removed, or changed with a field-level diff Keeping the shape written down — in a schema, a type, or a fixture — means a change to the shape is a visible diff rather than a silent runtime shift, and it means the integration test can assert against a concrete shape instead of guessing.

Configuration and code are kept apart: the configuration says what data comparison should do, the modules say how. A reviewer can change a threshold or a policy without reading the implementation, which keeps the risky changes small and lets the diff stand as the review.

The failure path through Row comparator is designed before the happy path is finished, because a floating-point field that differs by rounding noise and produces a huge, meaningless change set is the case that distinguishes a data comparison system from a demo. Every component declares what it does when its inputs are wrong, and the integration test asserts each of those declarations.

Prerequisites and dependencies

Have these ready before you install anything, so the data comparison setup is not blocked by a missing tool:

• Node.js 20 LTS or later
• Familiarity with data comparison fundamentals
• A local environment to run the test suite

The runtime dependencies are chosen to keep the data comparison datapath legible:

• commander (^12.1.0) — CLI framework
• csv-parse (^5.6.0) — Snapshot parsing

# setup for data-diff-cli
pnpm init
pnpm add commander csv-parse
pnpm add -D typescript tsx vitest

Versions are pinned to a major range so the data comparison install is reproducible; commit the lockfile. The snapshot reader reads its operational parameters from the environment and validates them at startup, so a missing value fails the boot rather than the first data comparison request.

Directory structure

Keep each data comparison component in its own file so it can be tested without starting the others:

data_diff_cli/
├── src/index.ts
├── src/snapshot-reader.ts
├── src/key-indexer.ts
├── tests/data_diff_cli.test.ts
└── package.json

Phase 1: Snapshot reader

Build the snapshot reader that anchors data comparison and defines the core data shape.

// Snapshot reader: data comparison core
export function snapshotreader(input: SnapshotreaderInput): SnapshotreaderResult {
  // Build the snapshot reader that anchors data comparison and defines the core data shape.
  return processSnapshotreader(input);
}

After this phase the snapshot reader works in isolation and is covered by a unit test. The snapshot reader defines the data shape that every later stage assumes, so getting it right first means the later phases compile against a stable contract.

Treat the snapshot reader as the single source of truth for the data comparison data shape. Validate inputs at the boundary and reject early, so a malformed value never reaches Key indexer or Row comparator where its effect would be harder to localize.

Phase 2: Key indexer

Add the key indexer so data comparison can computing the rows that correctly.

// Key indexer — handles data comparison Add the key indexer so
export class Keyindexer {
  constructor(private opts: KeyindexerOptions) {}
  async run(ctx: Context): Promise<KeyindexerOutput> {
    return this.execute(ctx);
  }
}

After this phase the key indexer works in isolation and is covered by a unit test. This is where data comparison correctness lives: the key indexer is what makes computing the rows that differ between two snapshots of a dataset and emitting a patch true, and its test is the proof.

Keep the key indexer pure where you can. A function that takes an input and returns an output is trivial to test and impossible to mock wrong; a class with hidden state is the opposite. Push side effects to the caller and keep the key indexer about the transformation.

Phase 3: Row comparator

Wire the row comparator to handle a floating-point field that without losing data.

// Row comparator: data comparison stage
export const rowcomparator = (next: Next) => async (req: Request) => {
  const out = await transform(req);
  return next(out);
};

After this phase the row comparator works in isolation and is covered by a unit test. The row comparator is the difference between a data comparison system that works on the happy path and one that survives a floating-point field that differs by rounding noise and produces a huge, meaningless change set.

Design the failure before the success. Decide what row comparator does when a floating-point field that differs by rounding noise and produces a huge, meaningless change set: reject, retry, shed, or degrade — and write that test first. A data comparison system whose failure behavior is unspecified will pick one at random under pressure, and it is usually the wrong one.

Phase 4: Patch emitter

Integrate the patch emitter to make data comparison observable and testable end to end.

// Patch emitter for data comparison
export async function patchemitter(stream: AsyncIterable<Chunk>): Promise<Aggregate> {
  let acc = init();
  for await (const c of stream) acc = merge(acc, c);
  return acc;
}

After this phase the patch emitter works in isolation and is covered by a unit test. Observability is not a finishing touch; it is how you know the data comparison system is doing what it claims in production.

Emit one structured event per meaningful action, with enough context to localize a failure without a second round-trip. The patch emitter is what turns “something is slow” into “the key indexer is slow for inputs matching this shape,” which is the difference between an alert and an investigation.

Phase 5: Summary reporter

Finalize the summary reporter so the data comparison system degrades gracefully under overload.

// Summary reporter: data comparison finalization
export function summaryreporter(state: State): Result {
  validate(state);
  return finalize(state);
}

After this phase the summary reporter works in isolation and is covered by a unit test. The summary reporter is what keeps overload from becoming corruption: when the data comparison system is asked for more than it can do, it degrades in a defined way instead of failing in an undefined one.

Pick a degradation strategy for summary reporter — shed, queue, or backpressure — and make it explicit in code and in a test. Leaving the behavior to chance is how a data comparison traffic spike becomes a data loss incident.

Testing

Each phase has a focused test: snapshot reader, key indexer, and row comparator are exercised in isolation, then an integration test drives the full data comparison pipeline and asserts the a floating-point field that case is handled.

Start with unit tests that pin each component’s contract: the snapshot reader returns the documented shape, the key indexer transforms correctly, and the row comparator fails in the documented way. These tests are fast and they localize regressions to the component that broke.

Then add an integration test that composes the components and drives a realistic data comparison scenario end to end. Assert the happy path first, then deliberately trigger a floating-point field that differs by rounding noise and produces a huge, meaningless change set and assert the system degrades rather than corrupts. Because the components are pure where practical, the integration test stays deterministic rather than flaky.

A data comparison suite must also prove the negative: bad input is rejected at the boundary, overload is shed rather than queued indefinitely, and a partial failure does not leave the system in an inconsistent state. These are the tests that catch the regressions a happy-path suite never would.

The entire data comparison suite runs with no external service; every dependency is either in-process or stubbed, so every case is reproducible in CI. That is what lets the row comparator failure-mode assertions live as tests rather than as a manual checklist that drifts out of date.

# run the data comparison test suite
pnpm vitest run

Deployment and operations

Run at least two data comparison instances behind a load balancer and share state in Redis. Set resource limits so data comparison overload surfaces as backpressure rather than failure.

Run at least two stateless instances behind a load balancer, and move any shared state into a shared store so the data comparison behavior is global rather than per instance. Size limits so the system sheds load before it runs out of memory, because a data comparison system that OOMs under load fails in a way that is hard to reason about.

Drain on deploy: stop accepting new work, let in-flight data comparison operations finish, then exit. Because the system holds no state that cannot be reconstructed, a rollout is a matter of cycling instances one at a time while watching the error rate.

When a dependency is unavailable, decide explicitly whether the data comparison system fails closed or degrades. Leaving that choice to chance is how a dependency outage becomes a silent data loss, so make the choice visible in code and in a test.

Security considerations

Validate all data comparison inputs at the boundary, reject oversized payloads before buffering them, and keep secrets in the environment rather than the repository.

The data comparison system is a choke point, so it is the right place to reject malformed input before it reaches key indexer or row comparator. Validate the shape and the size at the edge, and let the inner components assume what they receive is well-formed.

Never log secrets or full payloads; log enough to localize a failure without recording data that should not be on disk. Treat any external call as untrusted and validate its response before using it, because a data comparison system that trusts its inputs blindly is one malformed response away from corruption.

Keep dependencies pinned and run an audit in CI. A transitive dependency with a known vulnerability is still a vulnerability in your data comparison system, and the lockfile is what makes the audit meaningful.

Observability

Emit one structured event per data comparison operation with the inputs’ shape, the outcome, and a correlation id. Track error rate and latency per component and alert on a step change rather than a single spike.

The correlation id is the thread that ties a user-visible symptom to the data comparison decision and to the component log entry for the same operation. Without it, an investigation starts from a timestamp and a guess.

Watch the row comparator failure rate separately from the overall error rate. A jump in row comparator failures usually means a floating-point field that differs by rounding noise and produces a huge, meaningless change set, not a bug in key indexer, and the two have different owners and different fixes.

Chart the share of data comparison operations that shed, queue, or succeed. A change in the mix points at the cause faster than an aggregate latency can, because each class has a different root.

Troubleshooting

A floating-point field that differs by rounding noise and produces a huge, meaningless change set. Reproduce with a minimal fixture, then add a guard test so the data comparison system rejects the condition loudly.

Performance regresses under data comparison load. Profile with the sampling analyzer, locate the hot path in key indexer, and add a targeted benchmark to lock the regression.

Keep the data comparison dashboards separate from dependency dashboards. When an operation fails, the correlation id lets you decide in seconds whether the fault is snapshot reader, key indexer, or row comparator, instead of guessing from a timeout.

Limitations and next steps

• Single-region; multi-region replication is out of scope.
• The snapshot reader is in-process and not horizontally shared by default.
• No auth layer; assume deployment behind an edge that enforces identity.

The data comparison system is single-region by design; extending to multi-region means making snapshot reader shareable across regions, which is a real change to the consistency model rather than a config flag. The row comparator is in-process, which is correct for one instance but must move to a shared store before scaling out.

There is no authentication or authorization layer; the system assumes it is deployed behind an edge that enforces identity. Adding one is straightforward but out of scope for this build, and doing it here would distract from the data comparison correctness that is the point.

The natural next step, once the data comparison pipeline is trusted in production, is to add a short-TTL cache for idempotent reads and a richer patch emitter that exports histograms rather than counters. Both are additive and neither changes the contracts this build establishes.