Re: st: RE: identifying observations where multiple conditions are met

[Nick Cox <njcoxstata@gmail.com>]

In addition to several useful comments in this thread, note that the
thread started earlier today in
http://www.stata.com/statalist/archive/2013-11/msg00512.html  asks an
essentially similar question.

If it's OK to be either 0 (always) or 1 (always) then a direct
evaluation of that is

bysort id (type) : gen OK = inlist(type, 0, 1) & (type[1] == type[_N])

after which

edit if !OK

or

list if !OK

identifies problems.
Nick
njcoxstata@gmail.com


On 13 November 2013 20:50, Sergiy Radyakin <serjradyakin@gmail.com> wrote:
> On Wed, Nov 13, 2013 at 3:42 PM, Michael McCulloch
> <mm@pinestreetfoundation.org> wrote:
>> Thanks Elizabeth, yes it's all -id-, where "sometimes type==0, sometimes type==1, that I'm looking for.
>
> which means sd(type)!=0 within an ID group, hence see my code.
> to see particular IDs: tabulate id if flag
> to drop all 'problematic' obs: drop if flag
> Sergiy
>
>>
>> The code you suggest makes sense, however at the -collapse- stage, the records disappear -id- in which the "sometimes 0 / sometimes 1" existed prior to -collapse-.
>>
>>
>>
>> Best wishes,
>> Michael McCulloch
>>
>> --
>> Pine Street Foundation, since 1989
>> 124 Pine Street | San Anselmo | California | 94960-2674
>> P: (415) 407-1357 | F: (206) 338-2391 | http://www.PineStreetFoundation.org
>>
>> On Nov 13, 2013, at 12:29 PM, Allred, Elizabeth wrote:
>>
>>> On further reading, I see... You're trying to find ids with "mixed messages," i.e., sometimes 0 sometimes 1.
>>>
>>> gen zero=1 if type==0
>>> gen one=1 if type==1
>>> sort id
>>> collapse (sum) zero one, by(id)
>>> l id if zero==1 & one==1
>>>
>>> -----Original Message-----
>>> From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Michael McCulloch
>>> Sent: Wednesday, November 13, 2013 3:14 PM
>>> To: statalist@hsphsun2.harvard.edu
>>> Subject: st: identifying observations where multiple conditions are met
>>>
>>> Hello,
>>> I have data of the structure:
>>>       id              (range 1 to 300)
>>>       type    (for each -id-, should equal only one of two values: 0 or 1).
>>>
>>> In manual inspection of the data, I see instances exist where, for the same -id-,
>>>       some records have type==0, and
>>>       other records have type==1.
>>>
>>> If I enter
>>>       . tab id if type==0 & type==1
>>> obviously "no observations" is returned.
>>>
>>> How would I write a command in which I can tabulate all -id- for which some observations exist where type==0, and others where type==1? (At data entry, -type- should have unique for each -id-.)
>>>
>>>
>>> Best wishes,
>>> Michael McCulloch
>>>
>>> --
>>> Pine Street Foundation, since 1989
>>> 124 Pine Street | San Anselmo | California | 94960-2674
>>> P: (415) 407-1357 | F: (206) 338-2391 | http://www.PineStreetFoundation.org
>>>
>>>
>>> *
>>> *   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/