svy-io: Python bindings for ReadStat

Python bindings for ReadStat — read and write SAS, Stata, and SPSS files

Read and write statistical software files (SAS, Stata, SPSS) in Python. Fast, Pythonic API backed by ReadStat with full metadata support.
Keywords

svy-io Python, read SAS Python, read Stata Python, read SPSS Python, statistical file formats Python, ReadStat Python, sas7bdat Python, dta Python, sav Python

svy-io provides fast, Pythonic I/O for statistical software file formats. It wraps the ReadStat C library to read and write SAS, Stata, and SPSS files with full metadata support.

Installation

pip install svy-io

Features

  • Fast — Native C library performance via ReadStat
  • Polars-native — Returns Polars DataFrames for modern data analysis
  • Full metadata — Preserves variable labels, value labels, and file attributes
  • Multiple formats — SAS (.sas7bdat, .xpt), Stata (.dta), SPSS (.sav, .zsav, .por)

Quick Start

import polars as pl
from svy_io.stata import read_dta, write_dta
from svy_io.spss import read_sav, write_sav
from svy_io.sas import read_sas, read_xpt, write_xpt

# Read files
df_stata, meta_stata = read_dta("data.dta")
df_spss, meta_spss = read_sav("data.sav")
df_sas, meta_sas = read_sas("data.sas7bdat")

# Write files
write_dta(df, "output.dta", var_labels={"age": "Age in years"})
write_sav(df, "output.sav", var_labels={"age": "Age in years"})
write_xpt(df, "output.xpt", label="My Dataset")

Supported Formats

Format Read Write Extensions
Stata ✅ ✅ .dta
SPSS ✅ ✅ .sav, .zsav, .por
SAS ✅ ✅* .sas7bdat, .xpt

*SAS write support is for XPT (transport) format only.

Documentation

Stata

Read and write .dta files with variable labels, value labels, and metadata.

Stata File I/O →

SPSS

Read and write .sav, .zsav, and .por files with user-defined missing values.

SPSS File I/O →

SAS

Read .sas7bdat and .xpt files. Write FDA-compliant XPT transport files.

SAS File I/O →

Why svy-io?

Feature svy-io pandas pyreadstat
Returns Polars DataFrame ✅ ❌ ❌
Full metadata support ✅ ❌ ✅
Value labels ✅ ❌ ✅
User-defined missing ✅ ❌ ✅
Write support ✅ Limited ✅
Zip archive support ✅ ❌ ❌

Common Patterns

Convert Between Formats

from svy_io.stata import read_dta
from svy_io.spss import write_sav

# Stata → SPSS
df, meta = read_dta("input.dta")
write_sav(df, "output.sav")

Preserve Metadata

from svy_io.stata import read_dta, write_dta

# Read with metadata
df, meta = read_dta("input.dta")

# Transform data
df_filtered = df.filter(pl.col("age") > 18)

# Extract labels for remaining columns
var_labels = {
    v['name']: v.get('label')
    for v in meta['vars']
    if v['name'] in df_filtered.columns
}

# Write with preserved metadata
write_dta(df_filtered, "output.dta", var_labels=var_labels)

Export to Multiple Formats

from svy_io.stata import write_dta
from svy_io.spss import write_sav
from svy_io.sas import write_xpt

# Export to all formats
write_dta(df, "data.dta")
write_sav(df, "data.sav")
write_xpt(df, "data.xpt")
df.write_parquet("data.parquet")
df.write_csv("data.csv")

Differences from Stata/SPSS I/O

Feature SAS Stata SPSS
File formats .sas7bdat, .xpt .dta .sav, .zsav, .por
Value labels Separate catalog file In file In file
Write support XPT only Full Full
Tagged missing .A–.Z .a–.z User-defined
Default temporal coercion False (SAS7BDAT), True (XPT) False True

Requirements

  • Python 3.9+
  • Polars
  • ReadStat (bundled)