Marple DB

Marple DB is a data lakehouse built for time series data from automotive, aerospace and manufacturing. It ingests raw measurement files through configurable datastreams with built-in plugins for CSV, HDF5, MDF, and TDMS, or customised Python plugins. Priority signals can be imported first, and real-time data can be pushed through an SDK or API. Data is stored across hot, cold, and archive tiers, with automatic lifecycle movement; hot storage uses PostgreSQL, while cold storage uses Apache Iceberg with Parquet files, allowing direct access through engines like Trino, Spark, or PyIceberg, as well as MATLAB and Python SDKs. The system is designed for large datasets with high frequencies (200 kHz+), many signals (200k+), and very large file sizes (25+ GB), and its ingestion pipeline is optimised for measurement-file workloads. A unified namespace standardises signal names, post-processing can apply transformations and metadata, and the resulting structured and open-format data can be used directly for analysis, AI/ML, and data-mining workflows.