Help reshaping a dataset from long to wide

[Doris Rivera]

Dear Statalist, I have a dataset in which each citing_year (within cited family) adds a new set of citing_firms not present in a previous year. I need your help for reshape the below dataset as wide in which each row is a cited_family - citing_year, but the trick is that the following rows (within a given cited_family) must to add the citing_firms in the previous years (here rows).

That is, cited_family==1000065 has 9 unique firms in 2003. So, these 9 firms will be in a single row (9 new variables).

In 2004, I need the same firms present in the row from 2003 plus the new firm from 2004. So, this new row will have in this case one more variable.

In 2005, I need all the previous firms (in this case would be the 10 previous: 9 from 2003 and 1 from 2004 firms) plus the new firms from 2005, all in a new row for cited_family – citing_year.

Can you help me with this?
Thanks in advance for your help!

* Example generated by -dataex-. To install: ssc install dataex
clear
input long(cited_han_id cited_family) int citing_year
  2456791 1000065 2003
  4063354 1000065 2003
106357619 1000065 2003
112493276 1000065 2003
112589240 1000065 2003
113008532 1000065 2003
113011155 1000065 2003
113011156 1000065 2003
113011157 1000065 2003
112589245 1000065 2004
105475384 1000065 2005
105889591 1000065 2005
105889594 1000065 2005
105889596 1000065 2005
105889597 1000065 2005
105895227 1000065 2005
105895228 1000065 2005
105895229 1000065 2005
105895230 1000065 2005
100951950 1000065 2011
100951952 1000065 2011
100951953 1000065 2011
  1981214 1001348 2004
106051498 1001348 2004
106051499 1001348 2004
106051500 1001348 2004
106051501 1001348 2004
125793616 1001348 2004
125797054 1001348 2004
125799278 1001348 2004
125804427 1001348 2004
113013746 1001348 2006
113013747 1001348 2006
113013748 1001348 2006
113013749 1001348 2006
100959356 1001348 2008
100959357 1001348 2008
100959358 1001348 2008
100959359 1001348 2008
105892621 1001348 2010
185353888 1001348 2012
185367070 1001348 2012
185367071 1001348 2012
185532401 1001348 2012
185606878 1001348 2012
  2433998 1001348 2014
118295562 1001348 2014
118295563 1001348 2014
139562940 1001348 2014
139562941 1001348 2014
  3131499 1001348 2017
101363461 1001348 2017
101363462 1001348 2017
101363463 1001348 2017
101363464 1001348 2017
end

[Ken Chui]

* Generate an index:
bysort cited_family: gen seq = _n
* Reshape
reshape wide cited_han_id, i(cited_family citing_year) j(seq)

* Move family and year to front:
order cited_family citing_year

* Fill out missing vertically:
foreach x of varlist cited_han_id*{
    replace `x' = `x'[_n-1] if cited_family == cited_family[_n-1] & `x' == .
}

[Doris Rivera]

Dear Ken Chui, thanks a lot. It perfectly worked!