Generate new date variable by ID for panel data

[Devin Parker]

Hello,
I am trying to identify specific sets of consecutive dates by individual over the course of several years. I can calculate the length of stay between two dates, but I would like to generate (and add new rows/observations) for the days in between date1 and date2.

For example:

ID START END DURATION
1 2013-11-27 2013-11-29 3

I would like to have an expanded dataset that can list specific days, from var START:

ID START END DURATION
1 2013-11-27 2013-11-29 3
1 2013-11-28 2013-11-29
1 2013-11-29 2013-11-29

The expand function will only give me:

ID START END DURATION
1 2013-11-27 2013-11-29 3
1 2013-11-27 2013-11-29 3
1 2013-11-27 2013-11-29 3

The goal is to identify days, consecutive "in-hospital" days - versus - single-visit days. My thinking is that if I can identify each unique "in-hospital" day, then I can categorise and calculate them appropriately.

Any suggestions or alternative approaches would be greatly appreciated!
Thank you.

[Nick Cox]

Google dm0068 Stata Journal

[Clyde Schechter]

Assuming that your date variables are true Stata internal format date variables, not just string variables that appear as dates to human eyes, you just have to add one last step to what you've already done:

by ID START, sort: replace START = START[1] + _n - 1 if _n > 1

[Andrew Musau]

You should provide a data example using the dataex command as the specific code details will depend on whether your date variable is a proper date variable and whether the variable "duration" is numeric. Assuming that these are the case:

expand DURATION, g(new)
bys ID START (new): replace START= START+_n-1 if new

Note: Crossed with #2 and #3.

[Devin Parker]

Incredible! Thank you! Worked perfectly.