Skip to contents

HARplus: Enhanced R Package for GEMPACK .har and .sl4 Files

Author Last UpdatedStar

CRAN GitHub Version License

Extensions:
GTAPViz

Update (19-Oct-2025)

  • Writing HAR files without limitations – The new save_har() function fully supports writing .HAR files with no size restrictions, allowing up to seven dimensions and approximately two million elements per chunk.
  • Shock calculation and HAR export – Introduced shock_calculate_uniform() and shock_calculate() to compute and export shock results directly into GEMPACK-compatible .HAR files, supporting dynamic multi-period calculations (e.g., ONEY, TWOY, THRY, etc.) for recursive-dynamic simulations.

Overview

HARplus is an R package designed to process and analyze .HAR and .SL4 files, making it easier for GEMPACK users and GTAP model researchers to handle large economic datasets. It simplifies the management of multiple experiment results, enabling faster and more efficient comparisons without complexity.

With HARplus, users can extract, restructure, and merge data seamlessly, ensuring compatibility across different tools. The processed data can be exported and used in R, Stata, Python, Julia, or any software that supports .txt, CSV, or Excel formats.

Key Features

  • Efficient Data Extraction – Supports selective header loading and optimized memory usage for handling large .HAR and .SL4 files.
  • Flexible Data Structuring – Extract variables by name or dimension patterns while ensuring consistency across multiple inputs.
  • Customizable Aggregation & Merging – Manage subtotals, merge datasets, and structure data dynamically.
  • Multiple Export Options – Output extracted data in CSV, Stata, RDS, and Excel formats with structured formatting.
  • Powerful HAR Writing – Includes save_har() for exporting datasets to GEMPACK .HAR format with full binary compliance.
  • Designed for GEMPACK – Ensures smooth integration with .HAR and .SL4 files while offering additional flexibility.
  • Ideal for GTAP Model Users – Built specifically to process and analyze GTAP model results efficiently.

How It Works

HARplus simplifies .HAR and .SL4 file processing. You can: - Load files and selectively extract headers. - Extract data by variable name or dimension patterns. - Group, merge, and restructure data with ease. - Pivot and export data into structured formats. - Filter subtotals and rename dimensions for clarity.

Installation

HARplus (version 1.1.2) can be installed directly in R using:

install.packages("HARplus")

While the latest HARplus (version 1.1.3) can be installed from my GitHub using:

devtools::install_github("Bodysbobb/HARplus")

Quick Guide to HARplus

All commands in this package have several options that allow users to play around with the data more freely and efficiently, not just import and get the data. For a complete guide on HARplus functions, check out the Vignette or GitHub Vignette

Below is a categorized reference of the main functions in HARplus:

Data Importing

  • load_harx() – Loads .HAR files with selective header extraction and structured metadata.
  • load_sl4x() – Loads .SL4 files, extracting variable names and dimension structures.

Data Extraction

  • get_data_by_var() – Extracts specific variables from .HAR or .SL4 datasets, supporting subtotal filtering and merging.
  • get_data_by_dims() – Extracts data based on dimension patterns, with options for merging and subtotal filtering.

Data Structure Summary & Comparison

Data Grouping & Processing

  • group_data_by_dims() – Groups extracted data by dimension priority, with support for automatic renaming and subtotal handling.
  • rename_dims() – Renames dimension names for consistency.

Data Transformation

Data Export

  • export_data() – Exports extracted data to CSV, Stata, TXT, RDS, or XLSX, with support for multi-sheet exports.

HAR Exporting (New in v1.1.1)

  • save_har() – Saves processed data frames or arrays into GEMPACK-compatible .HAR files, automatically generating 1C set headers and supporting up to seven dimensions.

Technical Specifications

  • Supports both 1C (string) and RE (real) headers
  • Automatically generates 1C dimension sets (e.g., REG, COMM, ENDW)
  • Accepts data frames or arrays with flexible column naming
  • Writes associated set headers when export_sets = TRUE
  • Maintains full GEMPACK binary structure with no size limitation
  • Supports up to seven dimensions and approximately 2 million elements per chunk
  • Allows dimension renaming and supports duplicate dimension names (e.g., COMMxREGxREG) during export (New in v1.1.2)

Shock Calculation Framework (New in v1.1.2)

These functions provide a complete workflow to calculate, structure, and export GEMPACK-compatible shock files directly from .HAR, .SL4, .CSV, or .XLSX datasets—eliminating the need for manual conversion when preparing dynamic simulation shocks.

  • shock_calculate_uniform() – Calculates uniform percentage shocks across all base rates and exports directly to GEMPACK .HAR format. Supports additive (+, -) and multiplicative (*, /) adjustments.
  • shock_calculate() – Computes target-based shocks by comparing initial and target datasets, automatically exporting the resulting shocks to .HAR files with dynamic timeline headers (e.g., ONEY, TWOY, THRY, etc.).
  • create_initial_config(), create_target_config(), and create_calc_config() – Define input sources, column mappings, and timeline periods for use in both uniform and target-based shock calculations.

License & Author

HARplus is released under the MIT License. See the full license.

Author:
Pattawee Puangchit
Ph.D. Candidate, Agricultural Economics
Purdue University
Research Assistant at GTAP

Acknowledgements

Acknowledgement is due to Maros Ivanic for his work on the HARr package, which served as the foundation for HARplus. This package would not have been possible without his contributions.

GTAPViz: An Extension of HARplus for Visualization

I have developed another package specifically for visualization, particularly for GTAP users: GTAPViz

GTAP Database

Sample data used in this vignette is obtained from the GTAPv7 model and utilizes publicly available data from the GTAP 9 database. For more details about the GTAP database and model, refer to the GTAP Database.