Back to R project for statistical computing
GSoC 2026

Translating R Filters to AWK Streams

R's standard file reading functions must ingest data into memory before any filtering can occur, creating a hard memory ceiling for large datasets. AWK evaluates records one line at a time, filtering data before it ever enters R's memory space. The awkreader package bridges these two worlds, letting R users express filter conditions in familiar R syntax while AWK handles the heavy lifting at the shell level. The existing codebase is ~80% complete for Mac/Unix. This project completes and delivers the full production-ready CRAN package by: (1) completing and verifying the full translation engine: mapping R filter expressions including comparisons, logical operators, %in%, grepl(), and is.na() to correct AWK syntax across all platforms. (2) auditing and restoring Mac/Unix correctness. (3) achieving full Windows/PC compatibility via robust shell quoting and Rtools detection. (4) formalizing the package with roxygen2 docs, a cross-platform CI test harness with regression locking, and CRAN submission, enabling R users to filter datasets on standard hardware without memory constraints.

Project details

Contributor

Akshat Maurya

Mentors

Not available

Technologies

Not listed in the archive