v.22.3Improvement

Enhance Schema Inference with Globs in File/S3/HDFS/URL Engines

Improve schema inference with globs in File/S3/HDFS/URL engines. Try to use the next path for schema inference in case of error. #34465 (Kruglov Pavel).
Enhanced schema inference for File, S3, HDFS, and URL table engines with support for glob patterns and fallback to alternative paths in case of errors.

Why it matters

This feature improves the robustness and flexibility of automatic schema detection when querying external data sources via File, S3, HDFS, or URL engines. By using glob patterns, users can specify multiple files or paths to infer the schema, and the system will attempt the next available path on errors, minimizing failures due to inaccessible or malformed files.

How to use it

When defining tables with File, S3, HDFS, or URL engines, users can specify paths using glob patterns to match multiple files or objects. If schema inference fails on one path, ClickHouse automatically tries the next path in the pattern until a valid schema is found.