Adding rows for specific unique identifiers in a dataset conditional on some characteristics

[Andreas Mueller]

Dear Stata Community,

I am relatively new to Stata (switched from R to Stata) and I am currently experiencing some difficulties in adding rows to my dataset. I did not find an easy solution to do it apart from adding rows at the end of a dataset which is not desired in my case.

My dataset consists of more than 300'000 rows and I want to add rows for a specific "identifier" if there are insufficient observations/rows available. An example: A firm is delisted 12 months after the effective date of an M&A transaction and thus, only 12 rows are available in my dataset. In my analysis, however, I want to add rows for each firm with insufficient "rows" and fill in additional information until a period of at least 36 months is covered for each firm (i.e. I want to implement a similar approach as outlined by Mitchell and Stanfford (2000, p.298), Managerial Decisions and Long-Term Stock Price Performance).

Is there an easy command in order to do this time efficiently?

Thank you in advance and kind regards
Andreas Mueller

[Nick Cox]

You can add additional observations with (e.g.) the expand command.

I don't have the book or paper you refer to, so I cannot advise on how to add the "additional information".

[Mike Lacy]

Another possibility would be -fillin-. You would first need a variable to sequentially number the available observations within firm, if you don't have one. With that, -fillin- does what I think you want.

by firm: gen byte seqnum = _n
summ seqnum // verify that seqnum has min = 1, max = 36
fillin firm seqnum  //
//
sort firm seqnum // apparently order matters to you
tabulate _fillin // document new observations added

Regards, Mike

[Roberto Ferrer]

Andreas,

Operationally, how do you plan to "fill in additional information until a period of at least 36 months is covered for each firm" ?

You might want to check -merge-. Depending on the case, a -merge- will add observations to your original (i.e. master) data set and bring in additional information. All in one step.
See -help append-, also.