Skip to Main Content

Data Analysis with Stata

Loops

Unless you are doing something for which a Stata command does not already exist, it is entirely possible to use Stata without ever coding a loop. That being said, judicious use of loops can make life easier for you. This tutorial introduces you to the use of loops.

Loops in Stata follow a strict syntax and adhere to the same general principles as in other programming languages.  The following are the three types of loops in Stata. 

The use of each is best demonstrated using simple examples.

foreach

foreach is used to loop through essentially a list of words. Load the example dataset auto.dta using the sysuse command:

sysuse auto, clear

Suppose you want to rename the variables price and mpg to price_78 and mpg_78 respectively. You could of course type the rename command as many times as you need (in this case it would be twice), but you can automate this task using a foreach loop:

foreach v in price mpg {
  rename `v' `v'_78
}

The utility of looping becomes more obvious if we suppose that you need to rename ALL the variables in this dataset. If there are 100 variables, entering the rename command 100 times quickly becomes tedious. Here's how to add the suffix "_78" to all the variables in this dataset using a loop:

foreach v in * {
  rename `v' `v'_78
}

forvalues

forvalues is used to loop over consecutive values. For example, suppose you want to count the number of observations/rows in auto.dta.

local counter = 0
local N = _N

forvalues i = 1 / `N' {
  local counter = `counter' + 1
}
display `counter'

Note: this is exactly what's done by the count command.

while

while executes the code within the braces as long as the expression that it's evaluating is true. Here's Example 2 redone using a while loop.

local counter = 0
local N = _N

while  `counter' < `N' {
  local counter = `counter' + 1
}


display `counter'