maxframe.liteframe.read_json#

maxframe.liteframe.read_json(path, *, lines=True, dtype=None, usecols=None, chunk_bytes='64M', compression='infer', head_bytes='100k', default_index_type=None, storage_options=None, merge_small_files=True, session=None, **kwargs)[source]#

Read a JSON file into a LiteFrame with Arrow dtypes.

By default reads line-delimited JSON (JSONL/NDJSON) which is the natural fit for chunked distributed I/O.

Parameters:
  • path (str) – File path or glob pattern. Supports local and cloud storage paths.

  • lines (bool, default True) – Read the file as line-delimited JSON (one JSON object per line).

  • dtype (dict, optional) – Column name to dtype mapping for explicit type specification.

  • usecols (list, optional) – Return a subset of columns.

  • chunk_bytes (str, default '64M') – Target chunk size for splitting files.

  • compression (str, default 'infer') – Compression type. Inferred from file extension if ‘infer’.

  • head_bytes (str, default '100k') – Bytes to read for schema inference.

  • default_index_type (str, optional) – Type of default index to generate.

  • storage_options (dict, optional) – Extra options for storage connection.

  • merge_small_files (bool, default True) – Merge small files into larger chunks.

  • session (session, optional) – Execution session for remote dtype inference.

Return type:

LiteFrame

See also

LiteFrame.to_json

Write a LiteFrame to a JSON file.

read_csv

Read a CSV file into a LiteFrame.