embedded-sql-analytics

Runs analytical SQL over local files and in-memory Python objects with the DuckDB embedded engine. Read Parquet, CSV, and JSON files straight from their paths, treat globbed file sets as one table, join file data against Pandas or Polars frames, compute window functions and aggregations, and optionally persist intermediate results to a local database file. Use when a dataset is too large to load fully into memory, when SQL states a join or window calculation more clearly than DataFrame operations, when many partitioned files must be queried together, or when features must be engineered for a machine-learning model. Not for high-rate row-by-row writes, not for serving a shared multi-user database, and not for touching any data source beyond what the user has asked you to analyze.

Version

v1.0.0

Size

14 KB

Updated

2026-10-10

Your client verifies the signature before reading. This explicit manual read is available on every plan. Free automatically loads five chosen skills; paid plans automatically load relevant skills across the library.