Re: st: Re: question about ids in a group

[Pat R <profanepolicy@gmail.com>]

Hi Nick,

I did try using the -unique- option, but it did not keep the IDs
consistent for people across years. So a person would get say ID 4 in
2009, but ID 5 in 2010. If I'm not wrong, that might be because it
breaks the tie arbitrarily?

Best,
Pat

On Sun, Mar 20, 2011 at 6:22 AM, Nick Cox <njcoxstata@gmail.com> wrote:
> I wonder if you missed the -unique- option within -egen, rank()-.
>
> Nick
>
> On Sun, Mar 20, 2011 at 3:41 AM, Pat R <profanepolicy@gmail.com> wrote:
>
>> Thank you Joseph, this is exactly what I wanted! Now to figure out how
>> exactly it works :)
>
>> On Sat, Mar 19, 2011 at 11:33 PM, Joseph Coveney <jcoveney@bigplanet.com> wrote:
>>>
>>> Pat R wrote:
>>>
>>> I have a panel dataset from which I'm trying to make a smaller dataset
>>> with new identifiers. My (new) dataset is organised this way right
>>> now:
>>>
>>> QID    |   Year  |   PersonID
>>>
>>>  1       |   2009  |    3
>>>  1       |   2009  |    4
>>>  2       |   2009  |    9
>>>  2       |   2009  |    10
>>>  2       |   2010  |    9
>>>  2       |   2010  |    10
>>>  2       |   2009  |    11
>>>
>>> I want to create new identifiers for each person (consistent across
>>> years), starting from 1, within each QID. I tried using the -egen
>>> group- function, but since I cannot combine that with -by-, it creates
>>> unique values which ignore the QID. I tried using -egen rank-, but
>>> that of course skips a rank if two IDs are the same. I'm sure there's
>>> a terribly easy way to do this, but I just can't see it. Can anyone
>>> help? (It's my first time asking a question, so I apologize if I've
>>> left out important information)
>>>
>>> --------------------------------------------------------------------------------
>>>
>>> If I understand you correctly, then it would be something like that illustrated
>>> below.  If that's not what you want, then write back to the list with a fourth
>>> column in you example that shows just what you want as the result.
>>>
>>> Joseph Coveney
>>>
>>> . version 11.1
>>>
>>> .
>>> . clear *
>>>
>>> . set more off
>>>
>>> .
>>> . input long (QID Year PersonID)
>>>
>>>              QID          Year      PersonID
>>>  1.  1 2009  3
>>>  2.  1 2009  4
>>>  3.  2 2009  9
>>>  4.  2 2009 10
>>>  5.  2 2010  9
>>>  6.  2 2010 10
>>>  7.  2 2009 11
>>>  8. end
>>>
>>> .
>>> . *
>>> . * Begin Here
>>> . *
>>> . bysort QID PersonID: generate long new_id = _n == 1
>>>
>>> . quietly by QID: replace new_id = sum(new_id)
>>>
>>> . list, noobs sepby(QID)
>>>
>>>  +--------------------------------+
>>>  | QID   Year   PersonID   new_id |
>>>  |--------------------------------|
>>>  |   1   2009          3        1 |
>>>  |   1   2009          4        2 |
>>>  |--------------------------------|
>>>  |   2   2010          9        1 |
>>>  |   2   2009          9        1 |
>>>  |   2   2009         10        2 |
>>>  |   2   2010         10        2 |
>>>  |   2   2009         11        3 |
>>>  +--------------------------------+
>>>
>>> . exit
>>>
>>> end of do-file
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/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/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/