Adding an missing identifier when it is known in another year

[Sharon Blauw]

Hello,

I have a database where the gvkey idenitfier is known for all firm year observations. But the cik identifier is not know for all firm year observations. For the table below the missing value should be 1750 based on the combination of gvkey and cik in other fiscal years. Is there a way that I could generate this missing value for all observations where it is missing but I know what it should be based on the gvkey-fyear combination in other rows?
Thanks!

cik	gvkey	fyear
1750	1004	2015
1750	1004	2016
1750	1004	2017
1750	1004	2018
?	1004	2019

[Ulrich Wohak]

I'm assuming this is panel data? Maybe this fixes your issue:

xtset gvkey fyear
bys gvkey (cik): replace cik = cik[_N] if missing(cik)

[Sharon Blauw]

Yes, it is panel data. I tried your code, but it says "(0 real changes made)" after running.

[Ulrich Wohak]

What does the "?" in your data set look like? Is it represented as missing (i.e. "." or an empty string)? You could post an example using -dataex-.

Assuming cik is constant across gvkey and numeric you could also try

bys gvkey: egen cik2 = min(cik)
replace cik = cik2
drop cik2

[Sharon Blauw]

The question mark represents an "." in Stata. But I tried your second suggestion and that seems to solve my issue, so thank you very much for you help!