Replace missing values in group data

[Jason Cruso]

Hi everyone,

Was looking for a solution to the following problem.

I have a panel data that has some missing values as shown below.

Year  HoustId MedSalary
2002 1   .
2002 1   .
2002 1   .
2002 1   .
2002 1   .
2002 1   .
2002 1   .
2002 1   .
2002 1   23
2002 1   23
2002 1   23
2002 1   23

How can I replace the missing values with the actual group values that has the result data as shown below?

Year  HoustId MedSalary
2002 1   23
2002 1   23
2002 1   23
2002 1   23
2002 1   23
2002 1   23
2002 1   23
2002 1   23
2002 1   23
2002 1   23
2002 1   23
2002 1   23

The data is sorted by Year, HousetId.

In short, I need the entire (Year HoustId) to have the same non-missing (actual) MedSalary.

Thanks in advance,

J

[Mike Lacy]

bysort Year HoustId (MedSalary): replace MedSalary = MedSalary[1]

In Stata, the missing value indicator . is a very large numeric value, and such values sort to the end. Therefore, within groups defined by Year and HoustID, and sorted within those groups by MedSalary, the first observation within the group (i.e. [1]) will hold the nonmissing value of MedSalary, if such is available.

The bysort command is one of the most powerful features of Stata, so checking out the PDF manual entry available via -help bysort- will pay off.

[Jason Cruso]

Thanks a ton Mike! The code works....