use simulated-date.dta, clearIntroduction
Dates in any statistical software (Stata, R) are complicated.
For example, a particular date could be coded as “4-5-2021”, or “5-4-2021” or “April 5, 2021” or “5APR2021”.
In addition to the multiplicity of possible formats it is also difficult to do calculations on these kinds of quantities e.g. “How many days have elapsed from Day A to Day B?”
To address these issues, Stata wants these dates to be encoded as a number, specifically the number of days since January 1, 1960. We then make sure to format these numbers as dates.
Get The Data
List And Describe The Data
We see that both date variables are formatted as strings
list
describe | startdate enddate |
|------------------------|
1. | 2019-01-01 2019-1-30 |
2. | 2019-02-15 2019-5-30 |
3. | 2019-03-01 2019-4-30 |
+------------------------+
Contains data from simulated-date.dta
Observations: 3
Variables: 2 6 Apr 2021 16:46
---------------------------------------------------------------------------------------
Variable Storage Display Value
name type format label Variable label
---------------------------------------------------------------------------------------
startdate str10 %10s startdate
enddate str9 %9s enddate
---------------------------------------------------------------------------------------
Sorted by: Create Date Variables
There are many date functions in Stata, to work with different kinds of data in different formats.
help dateshould direct you to the documentation for Date and time functions.
generate startdate2 = date(startdate, "YMD") // create a date, specifying order of elements
generate enddate2 = date(enddate, "YMD") // create a date, specifying order of elementsThe command has created 2 dates in numeric form, but they display as numbers.
describe
listContains data from simulated-date.dta
Observations: 3
Variables: 4 6 Apr 2021 16:46
---------------------------------------------------------------------------------------
Variable Storage Display Value
name type format label Variable label
---------------------------------------------------------------------------------------
startdate str10 %10s startdate
enddate str9 %9s enddate
startdate2 float %9.0g
enddate2 float %9.0g
---------------------------------------------------------------------------------------
Sorted by:
Note: Dataset has changed since last saved.
+----------------------------------------------+
| startdate enddate startd~2 enddate2 |
|----------------------------------------------|
1. | 2019-01-01 2019-1-30 21550 21579 |
2. | 2019-02-15 2019-5-30 21595 21699 |
3. | 2019-03-01 2019-4-30 21609 21669 |
+----------------------------------------------+Format As Dates
We format these numeric variables as dates. While the variables remain as the number of days since January 1, 1960, since they are formatted as dates, they now appear as human readable dates.
format %d startdate2 enddate2
describe
listContains data from simulated-date.dta
Observations: 3
Variables: 4 6 Apr 2021 16:46
---------------------------------------------------------------------------------------
Variable Storage Display Value
name type format label Variable label
---------------------------------------------------------------------------------------
startdate str10 %10s startdate
enddate str9 %9s enddate
startdate2 float %d
enddate2 float %d
---------------------------------------------------------------------------------------
Sorted by:
Note: Dataset has changed since last saved.
+------------------------------------------------+
| startdate enddate startda~2 enddate2 |
|------------------------------------------------|
1. | 2019-01-01 2019-1-30 01jan2019 30jan2019 |
2. | 2019-02-15 2019-5-30 15feb2019 30may2019 |
3. | 2019-03-01 2019-4-30 01mar2019 30apr2019 |
+------------------------------------------------+Calculations
We can now use dates in calculations. For example, “How much time has elapsed between startdate2 and enddate2?”
generate elapseddays = enddate2 - startdate2
generate elapsedyears = (enddate2 - startdate2)/365
list, abbreviate(15) // list out the data with new variables | startdate enddate startdate2 enddate2 elapseddays elapsedyears |
|------------------------------------------------------------------------------|
1. | 2019-01-01 2019-1-30 01jan2019 30jan2019 29 .0794521 |
2. | 2019-02-15 2019-5-30 15feb2019 30may2019 104 .2849315 |
3. | 2019-03-01 2019-4-30 01mar2019 30apr2019 60 .1643836 |
+------------------------------------------------------------------------------+