Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: RE: marksample


From   "Ashim Kapoor" <ashimkapoor@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: RE: marksample
Date   Thu, 11 Sep 2008 18:00:33 +0530

Hello Nick

I believe sorting by a key simply amounts to the following:-

Consider row1 and row2 of a matrix. Now sort the matrix accordiing to
row1 which is your key. and we are done! sortrows works by sorting the
ONE and only row. All I need to do is add another key row and sort by
that row.

Am I wrong ?

I am reading mark from the programming manual. I get the idea of
st_views. I do not understand what the idea behind mark is. Why is
this variable named in a funny way like __0000 ? Is this 0 / 1 in each
row depending on some criteria ? What is that criteria?


Thank you,
Ashim.

On 9/11/08, Nick Cox <n.j.cox@durham.ac.uk> wrote:
> Your understanding is not correct.
>
> In your example of st_view()-, -touse- contains the name of one
> variable. This is explained in the help.
>
> Nor does -marksample- produce a list of variable names. It produces a
> single variable containing 1s or 0s as appropriate. As already explained
> in my previous message, it marks observations.
>
> I continue to think your strategy here ill-advised. You would need to
> write a special Mata program or think up special Stata code for almost
> every row-wise task you envisage.
>
> Nick
> n.j.cox@durham.ac.uk
>
> Ashim Kapoor
>
> Hello Nick,
>
> Thank  you for your time. Also thank you for your note on sortrows.
> Currently I am working on making " sortkeyrows" an ado file which will
> sort using a key. I will email the ado file to the list once I am
> done. I came across marksample while examining the code for sortrows.
>
> Let me rephrase my question.
>
> In the ado file for sortrows. we say " marksample touse".
> :
> :
> Later on we say st_view(X,., tokens(varlist), touse)
>
> Now if I understand st_view correctly touse should be a string of
> variable names which should be non zero for selection of values. But
> marksample is NOT producing is list of variable names.
>
> Where am I wrong?
>
> On 9/11/08, Nick Cox <n.j.cox@durham.ac.uk> wrote:
> > -marksample- is not part of Mata. It is a Stata command.
> >
> > It marks observations, not variables.
> >
> > It produces a new temporary variable, not a matrix.
> >
> > It is possible to use -marksample- independently of -syntax-:
> >
> > . sysuse auto
> > (1978 Automobile Data)
> >
> > . local if "2 == 2"
> >
> > . marksample touse, novarlist
> >
> > . su `touse'
> >
> >    Variable |       Obs        Mean    Std. Dev.       Min        Max
> > -------------+--------------------------------------------------------
> >    __000000 |        74           1           0          1          1
> >
> > However, it is usual to use -marksample- after -syntax-, or after a
> set
> > of commands that produced similar results. I can't think up a
> different
> > kind of situation in which -marksample- is the command of choice.
> >
> > Nick
> > n.j.cox@durham.ac.uk
> >
> > Ashim Kapoor
> >
> > I am trying to learn MATA. I have a query. What is marksample ? Does
> > it  return a matrix  with 0's in place of variables which are NOT
> > marked. Is it ONLY used after syntax ?
> >
> > I did not find the help very useful. Can someone please shed some
> > light on this ?
>
> *
> *   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index