pyspssio.read_sav

pyspssio.read_sav(spss_file, row_offset=0, row_limit=None, usecols=None, convert_datetimes=True, include_user_missing=True, chunksize=None, locale=None, string_nan='', data_only=False)[source]

Read data and metadata from SPSS file

Parameters:

  • spss_file (str) – SPSS filename (.sav or .zsav)

  • row_offset (int (default: 0)) – Number of rows to skip

  • row_limit (int (default: None)) – Maximum number of rows to return

  • usecols (Union[list, tuple, str, callable, None] (default: None)) – Columns to use (None for all columns)

  • convert_datetimes (bool (default: True)) – Convert SPSS datetimes to Python/Pandas datetime columns; False returns seconds from October 15, 1582 (SPSS start date)

  • include_user_missing (bool (default: True)) – Whether to keep user missing values or replace them with NaN (numeric) and “” (strings)

  • chunksize (int (default: None)) – Number of rows to return per chunk

  • locale (str (default: None)) – Locale to use when I/O module is operating in codepage mode

  • string_nan (Any (default: '')) – Value to return for empty strings

  • data_only (bool (default: False)) – Default (False) returns tuple of dataframe, metadata; If True, only return dataframe; Not applicable when chunksize is specified (always returns dataframe generator)

Return type:

Union[DataFrame, Tuple[DataFrame, dict], Generator[DataFrame, None, None]]

Returns:

  • tuple – DataFrame, metadata

  • generator – DataFrame(s) with chunksize number of rows (only if chunksize is specified)

Examples

Read data and metadata:

df, meta = pyspssio.read_sav("spss_file.sav")

Read metadata only:

meta = pyspssio.read_metadata("spss_file.sav")

Read data in chunks of chunksize (number of rows/records):

for df in pyspssio.read_sav("spss_file.sav", chunksize=1000):
    # do something