Skip to content

Configuration reference

All parameters below are keyword-only after path (where path is accepted).

read_dir / CsvDir / CsvChunksDir

Parameter Type Default Description
path str \| None None Directory to scan ("." when None)
extension str "csv" File extension without dot
delimiter str "," Field delimiter
encoding str "utf-8" Preferred text encoding
newline str "" Newline argument to open()
quotechar str '"' Fallback quote character
escapechar str \| None None CSV escape character
chunksize int \| None None If set and ≥ 1, return CsvChunksDir; omit or use chunking factories only when size is positive (invalid values raise ValueError).
strict_headers bool False Pin column set from the first sorted file for that iteration only; does not mutate expected_headers. Order of columns across files is ignored (Headers).
expected_headers list[str] \| None None Explicit required columns
on_mismatch "error" \| "skip" "error" Header mismatch policy
recurse bool False Walk subdirectories
case_insensitive bool True Case-insensitive extension match
include_hidden bool False Include dotfiles and hidden dirs

read_dir_chunks

Same as above except:

  • chunksize must be ≥ 1 (read_dir_chunks default is 1000; values less than 1 raise ValueError)
  • Always returns CsvChunksDir

CsvDirFile

Parameter Type Default Description
path str \| None None Directory to scan
extension str "csv" File extension
delimiter str "," Field delimiter
quotechar str '"' Quote character hint
escapechar str \| None None Escape character
encoding str "utf-8" Preferred encoding
newline str "" Newline for open()
strict_headers bool False Canonical header sequence: first scanned file after sorting when set, unless expected_headers is given (Headers); file naming determines which path sorts first
expected_headers list[str] \| None None Canonical header sequence (names and order)
on_mismatch str "error" "error" or "skip"
recurse bool False Recursive discovery
case_insensitive bool True Extension matching
include_hidden bool False Hidden files

Iterator method matrix

Row mode (CsvDir)

Method Yields
iter(reader) dict[str, str]
.with_names() / .enumerate() (stem, dict)
.with_paths() (path, dict)
.with_names().iter_column(c) (label, str)
.with_names().select_columns(cols) (label, dict)

Chunk mode (CsvChunksDir)

Method Yields
iter(reader) list[dict]
.enumerate() (stem, list[dict])
.with_paths() (path, list[dict])
.enumerate().iter_column_chunks(c) (label, list[str])
.enumerate().select_columns_chunks(cols) (label, list[dict])