pathlib.Path.glob() - ** vs *

Question

Accepted Answer

In glob patterns, a single * matches any sequence of characters that doesn't include a path separator. So '*.csv' matches 'sales.csv' but not 'reports/sales.csv'. The double ** is special: it matches zero or more directories. '**/*.csv' therefore matches 'sales.csv', 'reports/sales.csv', 'reports/q1/sales.csv', and so on at any depth. A subtle point: '**/*.csv' requires at least one path segment before the filename because of the slash after **. To also match files directly in /data, you need to combine both: itertools.chain(p.glob('*.csv'), p.glob('**/*.csv')), or simply use p.rglob('*.csv') which is shorthand for p.glob('**/*.csv') and does include the root level. In practice, rglob('*.csv') is the idiomatic way to find all CSVs under a directory tree, and it's the kind of one-liner that shows up constantly in data engineering scripts.