Re: st: strange error with stata10 and 9 using by

[Nick Cox <njcoxstata@gmail.com>]

In some lines here you refer to -fechalta- and in some to -fechaalta-.
But my guess is that you would see different messages if that was the
problem.

This works as expected:

. sysuse auto, clear
(1978 Automobile Data)

. sort foreign

. set seed 2803

. gen y1 = runiform()

. gen y2 = runiform()

. by foreign : gen flag = 1 if y1 > y2
(40 missing values generated)

. drop if flag
(74 observations deleted)

.
. sysuse auto, clear
(1978 Automobile Data)

. sort foreign

. set seed 2803

. gen y1 = ceil(10 * runiform())

. gen y2 = ceil(10 * runiform())

. by foreign : gen flag = 1 if y1 == y2
(70 missing values generated)

. drop if flag == 1
(4 observations deleted)

On Mon, May 23, 2011 at 3:24 AM,  <nicostat@gmail.com> wrote:

> I have a very strange error when I run the do file with stata 10 and 9
>  for example see the code below . The first variable is generate the
> second one not, if I eliminate" by codind"
> still to work. I don't understand what happen.
> thanks in advance
> best
> Nico
>
>  sort codind fechalta fechabaja
> by codind : gen coin=1 if fechalta>fechabaja
> drop if coin==1
> (64329 observations deleted)
> by codind : gen coin1=1 if fechalta==fechabaja
> drop if coin1==1
> coin1 not found
> r(111);
>
> gen coin1=1 if fechaalta== fechabaja
> drop if coin1==1
> (300 observations deleted)
> drop coin1

*
*   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/