Re: st: Sorting data in deciles and then regressing and storing coefficients. (Looping)

[Jeph Herrin <stata@spandrel.net>]

If you want one portfolio variable (rather than different ones for different years) you may want to simplify:

gen portfolio=.
levelsof year, local(years)
foreach Y in `years' {
         xtile tmp=cond(year==`Y',size,.), nq(10)
         replace portfolio=tmp if year==`Y'
	 drop tmp
}


J

On 7/11/2013 6:16 AM, C. Evans wrote:
> Please see below for my problem which has now been solved.
> 1. I am using -xtile-. Typing -which xtile- in Stata shows me:
> *! version 3.1.6  04feb2013
> Yet typing -which _gxtile- gives me:
> *! _gxtile version 1.2 UK 08 Mai 2006
>
> What would be the difference here as I don't know.
>
> 3. The exact command I typed to set up my portfolios. Firstly using J's suggested command, then using -rowtotal-
>
> -levelsof year, local(years)-
> -foreach Y in `years' {
>         xtile portfolio_`Y'=cond(year==`Y',size,.), nq(10)
> }-
> -egen portfolio = rowtotal( portfolio_1964 portfolio_1965 portfolio_1966 portfolio_1967 portfolio_1968 portfolio_1969
> portfolio_1970 portfolio_1971 portfolio_1972 portfolio_1973 portfolio_1974 portfolio_1975 portfolio_1976 portfolio_1977
> portfolio_1978 portfolio_1979 portfolio_1980 portfolio_1981 portfolio_1982 portfolio_1983 portfolio_1984 portfolio_1985
> portfolio_1986 portfolio_1987 portfolio_1988 portfolio_1989 portfolio_1990 portfolio_1991 portfolio_1992 portfolio_1993
> portfolio_1994 portfolio_1995 portfolio_1996 portfolio_1997 )-
> -drop portfolio_1964 portfolio_1965 portfolio_1966 portfolio_1967 portfolio_1968 portfolio_1969 portfolio_1970
> portfolio_1971 portfolio_1972 portfolio_1973 portfolio_1974 portfolio_1975 portfolio_1976 portfolio_1977 portfolio_1978
> portfolio_1979 portfolio_1980 portfolio_1981 portfolio_1982 portfolio_1983 portfolio_1984 portfolio_1985 portfolio_1986
> portfolio_1987 portfolio_1988 portfolio_1989 portfolio_1990 portfolio_1991 portfolio_1992 portfolio_1993 portfolio_1994
> portfolio_1995 portfolio_1996 portfolio_1997-
>
> This was because the command created portfolios for all of the years in my dataset which was 1964 to 1997. I just wanted
> one variable to display all of the portfolio numbering and as such used -rowtotal- and then dropped the other portfolio
> variables.
>
> Thanks again for all the help with this problem,
> Chris
>
> On Jul 10 2013, Nick Cox wrote:
>
> > 1. The version of -xtile()- on SSC is
> >
> > *! _gxtile version 1.2 UK 08 Mai 2006
> >
> > Your version statement must refer to something else.
> >
> > 2. -rowtotal()- is part of -egen-, i.e. part of official Stata.
> >
> > 3. The command
> >
> > egen portfolio = rowtotal()
> >
> > isn't legal. Please say what you did use to get better advice.
> >
> > Nick
> > njcoxstata@gmail.com
> >
> > On 10 July 2013 18:06, C. Evans <ce306@cam.ac.uk> wrote:
> > > Jeph, thanks for the suggestion it solved my problem (suggestion can be seen below). My version of -xtile()- which
> > > can be found in -egenmore- SSC. Is : version 3.1.6 04feb2013 . I'm using Stata 12.1. My aim was to create 10
> > > 'portfolios' a year based on -size- . Jeph's code and then -rowtotal()- solves this.
> > >
> > > The output that the command produced is below. I then used:
> > >
> > > egen portfolio = rowtotal()
> > >
> > > which is also from -egenmore- (SSC). -sample()- as I suggested earlier was not needed. -sample()- was just to reduce
> > > the number of observations by drawing a random sample from my dataset and using that instead of the full dataset.
> > >
> > >
> > >          id year size por~1964 por~1965 por~1966 por~1967.... |
> > >
> > > |----------------------------------------------------------------------|
> > >    1. | 10006 1964 219303.1 8 . . . |
> > >    2. | 10006 1965 249306.1 . 8 . . |
> > >    3. | 10006 1966 272942.9 . . 8 . |
> > >    4. | 10006 1967 277828.9 . . . 7 |
> > >    5. | 10006 1968 291790.6 . . . . |
> > >
> > > |----------------------------------------------------------------------|
> > >    6. | 10006 1969 289380.5 . . . . |
> > >    7. | 10006 1970 244459.8 . . . . |
> > >    8. | 10006 1971 297727.2 . . . . |
> > >    9. | 10006 1972 267790.8 . . . . |
> > >   10. | 10006 1973 256238.1 . . . . |
> > >
> > > |----------------------------------------------------------------------|
> > >   11. | 10006 1974 240341.8 . . . . |
> > >   12. | 10006 1975 230729.7 . . . . |
> > >   13. | 10006 1976 285526.4 . . . . |
> > >   14. | 10006 1977 304096.8 . . . . |
> > >   15. | 10006 1978 291149.2 . . . . |
> > >
> > > Thanks for all of the help,
> > > Chris
> *
> *   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/