Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: st: RE: loops question


From   "Martin Weiss" <[email protected]>
To   <[email protected]>
Subject   RE: st: RE: loops question
Date   Tue, 13 Jan 2009 14:03:24 +0100

Line for the server...

Are you positively sure the evaluation order is correct in your -drop-
statement (see [U]13.3)? Use braces where necessary! Note the different
number of dropped observations in

**********
sysuse auto, clear
drop if rep78==4 | f==1 & weight>2500
sysuse auto, clear
drop if (rep78==4 | f==1) & weight>2500
**********

HTH
Martin

-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Richard Harvey
Sent: Tuesday, January 13, 2009 12:42 PM
To: [email protected]
Subject: Re: st: RE: loops question

hi,

True.. Give that i am forming groups by id and year, that loop is not
required. I could just say grp=group(id year).

But once the grp had been defined and i run the following code..

levelsof grp, local(m)
foreach d of local m {
summ year if grp==`d'
local a=`r(mean)'
local b=`a'+1
di `a' "  " `b'
drop if month<m(`a'm4) | month>m(`b'm4) & grp==`d'
}

the drop if month<m(`a'm4) | month>m(`b'm4) & grp==`d' should be run
for each grp seperately should it not?

thanks
rich



2009/1/13 Martin Weiss <[email protected]>:
> Line for the server...
>
> This is much clearer and the story is complete now. It does not sound like
a
> standard task to me. I can spot a problem in
>
> sort id year month
> levelsof id,local(l)
> foreach c of local l {
> egen grp=group(id year)
> }
>
> You have at least two levels of id so the -foreach- would cycle at least
> twice, but after the first run you already have a variable "grp". So the
> second run through the loop should end with "grp already defined r(110)"?
> Can you confirm that?(If you do not condition on c in the loop, what is
the
> loop good for anyway?)
>
> HTH
> Martin
>
>
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]] On Behalf Of Richard Harvey
> Sent: Tuesday, January 13, 2009 12:06 PM
> To: [email protected]
> Subject: Re: st: RE: loops question
>
> hi martin,
>
> I am not sure what you mean by this " So far, as far as I can tell,
> you have only told the list half the story..."
>
> I am sorry if i was not clear earlier. I will try to rephrase it
>
> The following is what i have
> ***************************
> id       year      month         var1
>
> 1       1980      1980m1       x
> 1       1980      1980m2       x
> .......................
> 1       1980      2006m12     x
>
> 1       1981     1980m1        x
> 1       1981     1980m1        x
> ........................
> 1        1981    2006m12      x
>
> 2       1980      1980m1       x
> 2       1980      1980m2       x
> .......................
> 2       1980      2006m12    x
>
> 2       1981     1980m1       x
> 2       1981     1980m1       x
> ........................
> 2        1981    2006m12     x
>
>
> The following is what i want
> **********************
> id       year      month      var1
>
> 1       1980      1980m4    x
> .......................
> 1       1980      1981m4    x
>
> 1       1981     1981m4     x
> 1       1981     1980m1     x
> ........................
> 1        1981    2006m12   x
>
> 2       1980      1980m1   x
> 2       1980      1980m2   x
> .......................
> 2       1980      2006m12  x
>
> 2       1981     1980m1    x
> 2       1981     1980m1    x
> ........................
> 2        1981    2006m12  x
>
> and the following is my code that attempts to achieve the above..
>
> ********start code**********
>
> * create grp as combination of id and year
>
> sort id year month
> levelsof id,local(l)
> foreach c of local l {
> egen grp=group(id year)
> }
>
> * loop over the groups to keep the relevant months
>
> levelsof grp, local(m)
> foreach d of local m {
> summ year if grp==`d'
> local a=`r(mean)'
> local b=`a'+1
> di `a' "  " `b'
> drop if month<m(`a'm4) | month>m(`b'm4) & grp==`d'
> }
>
> *******end code************
>
> and that gives me the following result
>
>      id   year     month   grp
>    1181   1980    1981m4     1
>    1181   1981    1981m4     2
>    1181   1981    1981m5     2
>    1181   1981    1981m6     2
>    1181   1981    1981m7     2
>    1181   1981    1981m8     2
>    1181   1981    1981m9     2
>    1181   1981   1981m10     2
>    1181   1981   1981m11     2
>    1181   1981   1981m12     2
>    1181   1981    1982m1     2
>    1181   1981    1982m2     2
>    1181   1981    1982m3     2
>    1181   1981    1982m4     2
>
> So it works fine for the second group but not for the first group.
>
> I am wondering why? I hope this is much better.
>
>
> rich
>
> p.s what does Line for the server... mean???
>
>
> 2009/1/13 Martin Weiss <[email protected]>:
>> Line for the server...
>>
>> " What am I doing wrong?"
>>
>> The answer to that depends on how the outcome of your proposed solution
>> differs from what you want to achieve.
>>
>> HTH
>> Martin
>>
>>
>> -----Original Message-----
>> From: [email protected]
>> [mailto:[email protected]] On Behalf Of Richard Harvey
>> Sent: Tuesday, January 13, 2009 11:33 AM
>> To: [email protected]
>> Subject: st: loops question
>>
>> Hi all,
>>
>> I have a dataset set up as follows
>>
>> id       year      month         var1
>>
>> 1       1980      1980m1       x
>> 1       1980      1980m2       x
>> .......................
>> 1       1980      2006m12     x
>>
>> 1       1981     1980m1        x
>> 1       1981     1980m1        x
>> ........................
>> 1        1981    2006m12      x
>>
>> 2       1980      1980m1       x
>> 2       1980      1980m2       x
>> .......................
>> 2       1980      2006m12    x
>>
>> 2       1981     1980m1       x
>> 2       1981     1980m1       x
>> ........................
>> 2        1981    2006m12     x
>>
>> for each id year combination i want to keep the months from april of
>> that year to april of next year.
>>
>> i.e for id 1 and year 1980 combination i want to keep months from
>> april 1980 to april 1981.
>> again for id 1 and year 1981 combination i want to keep months from
>> april 1981 to april 1982.
>>
>> then for,
>>
>> id 2 and year 1980 combination i want to keep months from april 1980
>> to april 1981.
>> again for id 2 and year 1981 combination i want to keep months from
>> april 1981 to april 1982
>>
>> so the finaldataset should look like
>>
>> id       year      month      var1
>>
>> 1       1980      1980m4    x
>> .......................
>> 1       1980      1981m4    x
>>
>> 1       1981     1981m4     x
>> 1       1981     1980m1     x
>> ........................
>> 1        1981    2006m12   x
>>
>> 2       1980      1980m1   x
>> 2       1980      1980m2   x
>> .......................
>> 2       1980      2006m12  x
>>
>> 2       1981     1980m1    x
>> 2       1981     1980m1    x
>> ........................
>> 2        1981    2006m12  x
>>
>>
>> My logic was first to create  a variable grp, where grp=group(id
>> year), then loop over the grps to keep the relevant months. What am i
>> doing wrong?
>>
>> ********start code**********
>>
>> * create grp as combination of id and year
>>
>> sort id year month
>> levelsof id,local(l)
>> foreach c of local l {
>> egen grp=group(id year)
>> }
>>
>> * loop over the groups to keep the relevant months
>>
>> levelsof grp, local(m)
>> foreach d of local m {
>> summ year if grp==`d'
>> local a=`r(mean)'
>> local b=`a'+1
>> di `a' "  " `b'
>> drop if month<m(`a'm4) | month>m(`b'm4) & grp==`d'
>> }
>>
>> *******end code************
>>
>> --
>> thanks for your time
>> rich
>> *
>> *   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/
>>
>
>
>
> --
> thanks for your time
> rich
> *
> *   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/
>



-- 
thanks for your time
rich
*
*   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 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index