xpt.rs

Overview

xpttools (github: xpt.rs) is a Rust library and command-line tool suite for reading and working with SAS XPORT (XPT) version 5 files. These files are the standard format for data exchange in clinical research and are essential for regulatory submissions to agencies like the FDA.

Features

• Library: Use as a Rust crate in your projects for programmatic access to XPT files
• CLI Tool: Command-line utilities for inspection and conversion of XPT files
• Multiple Tools: Includes xptcols, xpthead, and xpt2csv for various operations
• Cross-platform: Works on macOS, Windows, and Linux
• Standards Compliant: Implements TS-140 specification for SAS Transport format

CLI Tools

The xpttools suite includes three main command-line utilities:

• xptcols: Print dataset metadata (variables, types, lengths, positions)
• xpthead: Display the first n rows of a dataset
• xpt2csv: Convert an XPT dataset to CSV format

CLI Usage Examples

./target/release/xpttools xptcols DM.xpt

./target/release/xpttools xpthead DM.xpt

./target/release/xpttools xpthead DM.xpt -n 20

./target/release/xpttools xpt2csv PC.xpt -o PC.csv

./target/release/xpttools xpt2csv SDTM.xpt -d PC -o PC.csv

./target/release/xpttools xpthead SDTM.xpt -d PC

./target/release/xpttools xpthead SDTM.xpt -d PC -n 5

Library Usage

Use xpt.rs as a Rust crate in your projects. For detailed library documentation and examples, see the USAGE.md documentation.

use xpttools::read_xpt_v5;

let datasets = read_xpt_v5("data.xpt")?;
let dataset = &datasets[0];
println!("Dataset: {} ({} variables, {} rows)", 
    dataset.name, dataset.vars.len(), dataset.rows.len());

use xpttools::read_xpt_v5_from_bytes;

let data = std::fs::read("data.xpt")?;
let datasets = read_xpt_v5_from_bytes(&data)?;

Technical Notes

• TS-140 Specification: Implements the official Record Layout of a SAS Version 5/6 Data Set in SAS Transport (XPORT) Format, including NAMESTR offsets, headers, and missing value rules
• IBM Hex Floating-Point: Correctly handles IBM/360 Hex Floating-Point format conversion to IEEE-754, including proper exponent bias handling
• Standards Compliance: The code implements the 80-byte card stream, NAMESTR (140-byte) layout, and IBM/360 (HFP) → IEEE-754 conversion, matching the SAS specification exactly

Use Cases