
Mock Data Generators for JSON Tests Demos and Load
Generate JSON test data with Faker, Hypothesis, SQL and jq: seeds for repeatable runs, collisions, and a property test that found real bugs.
Clients, formatters, linters and CLIs worth installing. Small tools that save hours.

Generate JSON test data with Faker, Hypothesis, SQL and jq: seeds for repeatable runs, collisions, and a property test that found real bugs.

Compare JSON documents in tests while ignoring ids and timestamps: strip keys, DeepDiff exclusions, placeholders that keep relations, jq del, float tolerance.

Snapshot tests for JSON in pytest with syrupy and Node's test runner: what a snapshot stores, why random ids break it, and matchers and scrubbing that fix it.

csvkit and Miller timed on the same 200,000-row file. Type inference makes csvkit's sort, stat and JSON commands 100 times slower and can rewrite your dates.

Turn a CSV into a browsable, searchable, filterable website with sqlite-utils and Datasette, with a JSON and CSV API, facets and full-text search built in.

Use Polars lazy scanning to filter and group a large CSV without loading it whole, set column types so zip codes stay text, and write Parquet as you go.

Define a block once with an anchor, reuse it with an alias, and merge it with the merge key. Real output shows overrides, sharing and where merging stops.

Block and flow styles for sequences and mappings, lists of mappings, empty collections and why a missing value is None, each with its parsed result.

The pipe keeps line breaks, the greater than sign folds them, and chomping indicators control the final newline. Six variants and the exact strings.

Plain, single quoted and double quoted YAML strings compared, with the characters that force quotes and a table of values that silently change type.

How YAML uses spaces to build structure, why a tab is rejected, and how one extra space turns a sibling into a child. Real parser output for each case.

Pull records out of a SOAP response with lxml: mixed-namespace children, xsi:nil, empty versus absent, repeated nodes and exact decimals.
We send one email when a new article is published. No spam.