Sometimes only parts of a dataset mean something to you. In this post, we show you how to subset a dataset in Stata by variables or by observations. We use the census.dta dataset installed with Stata as the sample data.
Subset by variables
* Load the data
sysuse census.dta
* See the information of the data
describe
Contains data from /Applications/Stata/ado/base/c/census.dta
obs: 50 1980 Census data by state
vars: 13 6 Apr 2014 15:43
size: 2,900
---------------------------------------------------------------------------
storage display value
variable name type format label variable label
---------------------------------------------------------------------------
state str14 %-14s State
state2 str2 %-2s Two-letter state abbreviation
region int %-8.0g cenreg Census region
pop long %12.0gc Population
poplt5 long %12.0gc Pop, < 5 year
pop5_17 long %12.0gc Pop, 5 to 17 years
pop18p long %12.0gc Pop, 18 and older
pop65p long %12.0gc Pop, 65 and older
popurban long %12.0gc Urban population
medage float %9.2f Median age
death long %12.0gc Number of deaths
marriage long %12.0gc Number of marriages
divorce long %12.0gc Number of divorces
----------------------------------------------------------------------------
Sorted by:
keep: keep variables or observations
There are 13 variables in this dataset. Say we would like to have a separate file containing only the list of the states with the region variable. We can use the keep command to do this.
keep state state2 region
describe
Contains data from /Applications/Stata/ado/base/c/census.dta
obs: 50 1980 Census data by state
vars: 3 6 Apr 2014 15:43
size: 900
---------------------------------------------------------------------------
storage display value
variable name type format label variable label
---------------------------------------------------------------------------
state str14 %-14s State
state2 str2 %-2s Two-letter state abbreviation
region int %-8.0g cenreg Census region
---------------------------------------------------------------------------
Sorted by:
Now only three variables remain in the data. Note that this change only applies to the copy of the data in memory, not the file on disk. You need to use the save command to make change to the file itself. You may want to be careful when you save this change, as you will permanently lose all the other variables that are not in the keep list. So here we save it as a new file called slist.
save slist
* Now if you load back the original file, you should still have all the variables
sysuse census.dta, clear
Note the clear option clears the current data in the memory, which contains the three variables we kept. Don't worry, you should still have it on your disk since we have saved it as slist.dta.
drop: drop variables or observations
The drop command also works in subsetting data. Say we only need to work with population of different age groups. We can remove other variables and save as a new file called census2.
drop medage death marriage divorce
describe
Contains data from /Applications/Stata/ado/base/c/census.dta
obs: 50 1980 Census data by state
vars: 9 6 Apr 2014 15:43
size: 2,100
---------------------------------------------------------------------------
storage display value
variable name type format label variable label
---------------------------------------------------------------------------
state str14 %-14s State
state2 str2 %-2s Two-letter state abbreviation
region int %-8.0g cenreg Census region
pop long %12.0gc Population
poplt5 long %12.0gc Pop, < 5 year
pop5_17 long %12.0gc Pop, 5 to 17 years
pop18p long %12.0gc Pop, 18 and older
pop65p long %12.0gc Pop, 65 and older
popurban long %12.0gc Urban population
----------------------------------------------------------------------------
Sorted by:
save census2
Subset by observations
We can also use keep and drop commands to subset data by keeping or eliminating observations that meet one or more conditions. For example, we can keep the states in the South.
* Load the data again and clear the current one in memory
sysuse census.dta, clear
* See the contents of region
tabulate region
Census |
region | Freq. Percent Cum.
------------+-----------------------------------
NE | 9 18.00 18.00
N Cntrl | 12 24.00 42.00
South | 16 32.00 74.00
West | 13 26.00 100.00
------------+-----------------------------------
Total | 50 100.00
Note region is an integer type variable with a value label called "cenreg" indicating the four regions. We can use label list to see which integers are associated with the region labels.
label list cenreg
cenreg:
1 NE
2 N Cntrl
3 South
4 West
* The states in the South are coded as 3.
* Keep the observations/rows (the states) that are in the South region
keep if region==3
(34 observations deleted)
* Here are the 16 South states (rows) remaining in the dataset.
tabulate region
Census |
region | Freq. Percent Cum.
------------+-----------------------------------
South | 16 100.00 100.00
------------+-----------------------------------
Total | 16 100.00
Now let's use drop to eliminate those states with a population below the average.
* Load back the data
sysuse census.dta, clear
* summary statistics, mean=4518149
summarize pop
Variable | Obs Mean Std. Dev. Min Max
-------------+---------------------------------------------------------
pop | 50 4518149 4715038 401851 2.37e+07
* drop the rows/states with population less than the mean
drop if pop < 4518149
(33 observations deleted)
* list the states remained (those with population above the average)
list state
+---------------+
| state |
|---------------|
1. | California |
2. | Florida |
3. | Georgia |
4. | Illinois |
5. | Indiana |
|---------------|
6. | Massachusetts |
7. | Michigan |
8. | Missouri |
9. | New Jersey |
10. | New York |
|---------------|
11. | N. Carolina |
12. | Ohio |
13. | Pennsylvania |
14. | Tennessee |
15. | Texas |
|---------------|
16. | Virginia |
17. | Wisconsin |
+---------------+
You may want to be careful when using the list command. In this case, the census.dta is a small dataset with only 50 rows/observations in it, and we eliminated 33 observations, so we know we have a small number of cases to be listed in the output. If you are working with a big dataset, you may not want to list too much information to your output.
References
- StataCorp. (2017). Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC.
- StataCorp. (2017). Stata 15 Base Reference Manual. College Station, TX: Stata Press.
Yun Tai
CLIR Postdoctoral Fellow
University of Virginia Library
October 14, 2016
Updated May 23, 2023
For questions or clarifications regarding this article, contact statlab@virginia.edu.
View the entire collection of UVA Library StatLab articles, or learn how to cite.