Re: st: Merging w/ substituting for years with missing values

["James Beard" <james@beard.net>]

Another interpretation of the question would be "how to do the 
merge". This can be acheived by using -expand- on the A dataset to 
(say) produce (a total of) 5 records for each original record, then 
merging this dataset with B on year and doing the -carryforward-.

* on dataset A
gen original_year = year
expand 5
sort original_year
by original_year: replace year = year - 3 + _n 

Then use -merge- and -carryforward- as appropriate and finally drop 
the records with year!=original_year. Because there may now be 
duplicates on year, the merge would probably need to be m:1.

On Fri, 02 Aug 2013 12:44:48 -0400, David Kantor wrote:

It's not too hard to write code to do that, but you may want to try 
carryforward from SSC. It goes forward, but you can use a negative 
date to trick it into gong backward. For filling inward, I like to do 
both forward and back, and see if the values agree.

HTH
--David

At 11:24 AM 8/2/2013, Leoni wrote:

I'm merging two panel datasets (A and B), one of which has few years 
(A, I only need those), while the other has observations for many 
years (B) -- but also has missing values for some of them. As most of 
the variables that have missing values don't change a lot, I want to 
tell STATA to use values from previous or following years (from B) if 
they're missing for the exact year I need (those in A). This sounds 
rather messy but is there an elegant way of doing this in STATA?  

Any help is much appreciated.

Leoni
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/