egenmisc

Faster implementations of several egen functions. These are drop-in replacements that produce identical results to the corresponding functions in egenmore but run significantly faster, especially with by() groups.

Functions

Function	Description
fastxtile	Categorize a variable into quantile groups
fastwpctile	Compute (weighted) percentiles
fastcorr	Sample correlation
fastcov	Sample covariance
fastsd	Sample standard deviation
fastvar	Sample variance

All functions support by() for group-specific calculations.

Examples

sysuse nlsw88, clear

* Quartiles of wage by race
egen wage_quartile = fastxtile(wage), by(race) nq(4)

* Median wage by industry
egen med_wage = fastwpctile(wage), by(industry) p(50)

* Standard deviation by occupation
egen sd_wage = fastsd(wage), by(occupation)

* Correlation between hours and wage by race
egen corr_hw = fastcorr(hours wage), by(race)

* Covariance with minimum observation requirement
egen cov_hw = fastcov(hours wage), by(race) min(10)

Installation

From SSC:

ssc install egenmisc

To install the latest version from Github (Stata 13+):

net install egenmisc, from("https://raw.githubusercontent.com/matthieugomez/egenmisc/main/")

For Stata 12 or older, download the zip file and install from a local folder:

net install egenmisc, from("path_to_folder")