ReadMeFeed / Community knowledge
Real questions. Useful conversations. Find the people who know your stack.
Databases & messaging · #duckdb
Choose partitioned Parquet writes from DuckDB when downstream reads filter on a low‑cardinality column. Otherwise a single file reduces overhead. This guide compares trade‑offs and shows validation steps.
DuckDB can read Parquet files directly with SQL, pushing filters into the file reader for ad-hoc analytics without an ETL load step.
DuckDB's columnar storage and vectorized execution engine are a deliberate engineering trade-off: optimize for analytical scan/aggregation throughput at the cost of point-lookup and transactional performance. Here's how to verify it works for your workload.
Query Parquet files on disk with DuckDB's read_parquet, using column projection and predicate pushdown to read only what your SQL needs.
A decision guide for choosing between DuckDB's native .db storage and Apache Parquet, focusing on ACID compliance, I/O performance, and data portability.
Deciding between Parquet files, native .duckdb storage, and in-memory tables? This guide compares performance, concurrency, and portability to help you choose the right DuckDB storage strategy.
Learn how to write query results to a Hive-style partitioned Parquet dataset using DuckDB's COPY TO, validate the output, and recover from common write failures.
DuckDB's read_parquet lets you query Parquet lakes without ETL, but per-query metadata overhead and schema drift have real costs. Here's how to decide.
DuckDB’s vectorized engine and Parquet push‑down cut CPU work and I/O for analytical scans, as shown in a step‑by‑step example that creates a 100‑million‑row Parquet file, runs a filtered aggregation, and verifies push‑down with EXPLAIN.