Re: st: RE: generate new variables using the first record of the group

[Nick Cox <njcoxstata@gmail.com>]

No answer to this as yet, but for reference here is one way of doing this:

I assume cities are labelled 1 to 42. For "any value of 42"

forval i == 1/42 {
      su city_n if city == `i', meanonly
      gen city`i'_n = r(min)
}

Still sounds a bad idea....
Nick
njcoxstata@gmail.com


On 30 October 2013 00:29, Nick Cox <njcoxstata@gmail.com> wrote:
> The same answer and the same question. To get your -city_1- and
> -city_2- it is sufficient (although not necessary) to use any value of
> -city- that is equal to 30 or 12 respectively.
>
> But why do you want to do this?
> Nick
> njcoxstata@gmail.com
>
>
> On 29 October 2013 23:53, Song, Lin <Lin.Song@kingcounty.gov> wrote:
>> Thanks Sergiy and Nick.  Sergiy told me a command that I did not know.
>>
>> On the other hand, I did not ask my question clearly.  I was actually thinking something like the question below:
>>
>> For the following dataset , I'd like to generate two new variables in the last two columns (city1_n and city2_n) as shown below based on the first three columns.  I thought I could use egen.
>>
>> id      city    city_n  city1_n city2_n
>> 1       1       30      30      12
>> 2       1       30      30      12
>> 3       1       30      30      12
>> 4       2       12      30      12
>> 5       2       12      30      12
>> 6       2       12      30      12
>> 7       2       12      30      12
>> 8       .               30      12
>> 9       .               30      12
>> 10      .               30      12
>>
>> Thanks.
>>
>> Lin
>>
>> -----Original Message-----
>> From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Nick Cox
>> Sent: Tuesday, October 29, 2013 4:25 PM
>> To: statalist@hsphsun2.harvard.edu
>> Subject: Re: st: RE: generate new variables using the first record of the group
>>
>> Sergiy's answer is perfect, naturally, but quite why you would want to do this remains open. For some graphical purposes, a constant variable can be useful, but otherwise a constant is often better stored in a scalar or local.
>>
>> Nick
>> njcoxstata@gmail.com
>>
>>
>> On 29 October 2013 22:49, Sergiy Radyakin <serjradyakin@gmail.com> wrote:
>>> Lin, use the square brackets to address the particular observation of
>>> a variable:
>>>
>>> sysuse auto
>>> generate newvar=price[7]
>>>
>>> Creates a variable newvar, which is equal to the 7th observation of
>>> existing variable price.
>>>
>>> Best, Sergiy Radyakin
>>>
>>>
>>> On Tue, Oct 29, 2013 at 6:02 PM, Song, Lin <Lin.Song@kingcounty.gov> wrote:
>>>> Dear Stata users?
>>>>
>>>> How do I generate a new variable with its value for all observations equals to the first observation or the nth observation of another variable?
>>>>
>>>> Thanks.
>>>>
>>>> Lin
>>>>
>>>>
>>>>
>>>>
>>>> *
>>>> *   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/
>> *
>> *   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/
*
*   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/