1 views collected around this technical thread.
The article benchmarks cjson, rapidjson, and yyjson JSON libraries on a 6‑core OpenResty setup, shows that combining rapidjson's encode with yyjson's decode reduces 500,000 mixed encode/decode operations to 11 seconds, and describes a Lua wrapper that adds schema validation and improves maintainability.
This article explains a systematic process for evaluating and selecting the most suitable high‑performance JSON library for Python, covering the need assessment, benchmark definition, filtering by additional requirements, and detailed benchmark results that highlight orjson as the fastest option for small‑message encoding while discussing trade‑offs such as safety, customizability, and ecosystem support.
The article investigates why Python's built‑in json.dumps can be slow on large data structures, compares the performance of ujson, python‑rapidjson, simplejson, and the Rust‑based orjson, provides benchmark code, installation commands, and discusses a bug in ujson's indent handling.