How to Convert YAML to JSON: Anchors, Multiline Strings, and the Norway Problem

YAML as a JSON Superset

YAML (YAML Ain't Markup Language) is a human-friendly data serialization language commonly used in DevOps, Kubernetes manifests, CI/CD pipelines (GitHub Actions, GitLab CI), and Docker Compose. Officially, YAML 1.2 is a superset of JSON: every valid JSON document is valid YAML, but YAML adds extensive features like indentation scoping, object anchors, multiline scalars, and implicit type conversion.

The Infamous "Norway Problem" in YAML

In YAML 1.1, country code abbreviations such as NO (Norway) or ON (Ontario) are automatically coerced to boolean false and true unless explicitly quoted in strings. When converted to JSON, this creates subtle bugs where country lists become ["US", "GB", false, "DE"].

# DANGEROUS in YAML 1.1:
countries:
  - US
  - NO  # Naively parsed as boolean false!

# CORRECT:
countries:
  - "US"
  - "NO" # Explicit string quotes prevent boolean coercion

Anchors (&) and Aliases (*): Resolving Object References

YAML allows developers to define reusable configuration blocks using anchors (&anchorName) and inject them elsewhere using aliases (*anchorName) or merge keys (<<: *anchorName). When converting YAML to JSON, these references must be fully dereferenced and cloned into independent JSON objects:

default_db: &db_config
  host: localhost
  port: 5432
  pool: 10

production:
  <<: *db_config
  host: db.prod.cosharex.internal
  pool: 50

Multiline Strings: Literal (|) vs Folded (>)

Converting YAML Safely in TypeScript & Python

import * as yaml from 'js-yaml';

export function yamlToJson(yamlContent: string): string {
  // Use FAILSAFE_SCHEMA or JSON_SCHEMA to avoid unintended type coercion
  const parsed = yaml.load(yamlContent, {
    schema: yaml.JSON_SCHEMA
  });
  return JSON.stringify(parsed, null, 2);
}

import yaml
import json

def convert_yaml_to_json(yaml_str: str) -> str:
    # Always use safe_load to prevent arbitrary code execution vulnerabilities
    data = yaml.safe_load(yaml_str)
    return json.dumps(data, indent=2)