RCALLCOUNTRYCODE: Call R's countrycode package from Stata using rcall

• Current version: 0.1.11 10mar2021
• Contents: updates description install usage to do author

---

Updates

• 0.1.11 10mar2021
  (Since 0.1.0):
  • bug fixes for encoding issues, diacritics
  • handles moving of data with R package haven
  • improved error checking and handling
  • adds codelist option to get available codes from countrycode
  • set default arguments for options from, to and gen
• 0.1.0 16feb2019:
  • first version of the command

Description

This command uses rcall to call R's countrycode. It is a substitute for the kountry package from SSC.

I'd like to thank the authors of both packages:

• countrycode was written by Vincent Arel-Bundock, Nils Enevoldsen, and CJ Yetman.
• rcall was written by E. F. Haghish

There are a few advantages to using rcallcountrycode relative to kountry:

• it gets the functionalities of countrycode for R, which has a broader and more up-to-date coverage on country names in different formats and languages
• it scales well with the size of the dataset. See the benchmark (Win7/Linux) for a comparison between rcallcountrycode and kountry.
  • Because it asks R to convert only the unique strings in the dataset (which should not exceed the number of countries in the world in most use cases), applying it to a dataset of 200 or 200000 observations makes little difference. The current version of kountry does not scale well in large datasets.
  • In the current benchmark (v0.1.5), I add repeated country names 2000 times to a list of 196 countries. rcallcountrycode, which is much slower in the small dataset, only takes around 10%/50% (Win7/Linux) longer to run in the larger dataset. kountry takes 1000+ times longer than it does in the small dataset.

Machine	Dataset	rcallcountrycode	kountry
Win 7, 4-core 3.60GHz, 32GB RAM	196 countries	4.09 sec	0.02 sec
Win 7, 4-core 3.60GHz, 32GB RAM	196 countries x 2000	4.54 sec	17.88 sec
Ubuntu, 2-core 2.20GHz, 16GB RAM	196 countries	1.44 sec	0.01 sec
Ubuntu, 2-core 2.20GHz, 16GB RAM	196 countries x 2000	2.19 sec	21.27 sec

The main disadvantage is that rcallcountrycode requires additional dependencies, while kountry can be run directly after installing it from SSC without any additional work.

Install

Easier

1. Install R first (see below how)
2. Install rcall with the method recommended by its author: install the github package for Stata and then install rcall:

net install github, from("https://haghish.github.io/github/") replace
gitget rcall

3. Install rcallcountrycode:

github install luispfonseca/stata-rcallcountrycode

These steps should take care of all the dependencies automatically.

Harder

1. Install R first (see below how)
2. Install this package:

cap ado uninstall rcallcountrycode
local github "https://raw.githubusercontent.com"
net install rcallcountrycode, from(`github'/luispfonseca/stata-rcallcountrycode/master/)

3. Make sure you install all the dependencies

Dependencies

For this command to work, you need the following:

R

You need to have R installed. You can download RStudio here, which will install R on your computer and give you a graphical interface.

If you are not using github install to install rcallcountrycode, you also need to install the countrycode and haven packages in R:

install.packages("countrycode")
install.packages("haven")

Stata

Install rcall following the instructions in the page. The following commands currently work:

net install github, from("https://haghish.github.io/github/") replace
gitget rcall

Some commands from gtools by Mauricio Caceres Bravo are used to speed up this command when available, but are not required. Follow the instructions in the link to install, especially if you are dealing with large datasets.

Usage

input str20 country
"portugal"
"united kingdom"
"france"
"italy"
"spain"
"germany"
"germany"
"italy"
"switzerland"
"curaçao"
"côte d'ivoire"
"namibia"
""
"not a real country"
end
compress

* standardize country names stored in a variable named country (both are equivalent)
rcallcountrycode country, gen(countryname_en)
rcallcountrycode country, from(country.name) to(country.name) gen(countryname_en)

* get the ISO2 country codes
rcallcountrycode country, from(country.name) to(iso2c) gen(iso2code)

* get the country names in german
rcallcountrycode country, from(country.name) to(country.name.de) gen(countryname_de)

* get list of available codes from R
rcallcountrycode codelist

To do:

• Provide better diagnostics for non-matches

Author

Luís Fonseca
London Business School
lfonseca london edu
https://luispfonseca.com