st: Re: working within bygroups

[Christopher F Baum <baum@bc.edu>]

To indicate whether a borrower's issues in a particular market are 
seasoned, Nick wrote

bysort borrower mkttype (issuedate) : gen newmkttype = _n > 1

On May 8, 2004, at 2:33 AM, Adrian wrote:


> Nick, thank you very much. I was not familiar with this structure of
> generating variables by assigning them a value and adding extra
> conditions, like this:
>
> _n > 1
>
> I think I still need time to digest this as I don't understand very 
> well
> the structure behind this command. It has certainly been very useful,
> though. Thank you.
Note that Nick's command

gen newmkttype = _n>1

is not "assigning a value and adding extra conditions": it is defining 
a Boolean
condition, evaluating to (0,1). Let us say that within a by-group 
(simple structure: issuer  issue_date)
we want to flag those issues which are the third issues by that issuer. 
Then

bysort issuer (isser_date) : gen thirdiss = _n==3

would do that. It is just a matter of setting a logical condition on 
the RHS which expresses
what you would like to pick off. The above can be written in terms of 
instructions to a RA:

For each one of the issuers,
    Sort the issues by issue_date
    Put a flag on the third issue of that issuer

and Adrian's original problem is just a variation on that theme (with 
one more by-variable: the market type).

Kit

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/