This page collects the most common issues encountered when working with DuckDB and links to the detailed documentation for each. If you run into a problem that is not covered here, please consult the relevant section of the documentation or the how-to guides.
File Lock Errors
DuckDB allows only a single process to write to a database file at a time. Errors such as “Could not set lock on file” usually mean another process (for example, an editor, a notebook kernel, or a second CLI session) still holds the database open. The concurrency model and how to work around it are described in the following page.
Out-of-Memory Errors
When a query needs more memory than is available, DuckDB spills intermediate results to disk, but some workloads can still exhaust memory. The following pages explain how to diagnose and resolve out-of-memory situations.
CSV Type Inference
The CSV reader detects column types and the dialect automatically. When a column is read with an unexpected type, or rows fail to parse, the following pages explain how to inspect and override the detection.
Extension Installation
Extensions are downloaded and loaded at runtime. Installation can fail in offline or restricted environments. The following pages cover installation and its troubleshooting.
S3 and Cloud Storage Access
Reading from and writing to S3 and S3-compatible storage requires the httpfs extension and correctly configured credentials. The following pages cover the S3 API and authentication.