Re: st: calculating cumulative values of other observations

[Eric Booth <eric.a.booth@gmail.com>]

<>


******************!

//input example data//
clear
inp ID	str18 date  str4 activity hours   	
1	20071230  A	10	   			
1	20071122  A	11	  
2	20071120  A	5	  
3	20071210  B	6	  
3	20071111  B	24	  
3	20071130  F	20	  
3	20070511  B	49	  
4	20070611  A	50	  
end

//cleanup//
g date2 = date(date, "YMD")
format date2 %td
sort ID activity date2

//gen prior hours var//
g priorhrs = 0
forval n = 1/`=_N' {
by ID activity: replace priorhrs =  ///
	priorhrs + hours[_n-`n']  ///
	if date2[_n]>=date2[_n-`n'] 
	}
  l
********************!


- Eric

__
Eric A. Booth
Public Policy Research Institute 
Texas A&M University
ebooth@ppri.tamu.edu
+979.845.6754


On Apr 6, 2012, at 6:09 PM, KOTa wrote:

> Hi all,
> 
> i have a data looking like this:
> 
> ID	date	            activity	hours prior	     result should be:	
> 1	20071230  	A	10	   ?	<--	11
> 1	20071122	        A	11	   ?		0
> 2	20071120	        A	5	   ?		0
> 3	20071210	        B	6	   ?		73
> 3	20071111	        B	24	   ?		49
> 3	20071130	        F	20	   ?		0
> 3	20070511	        B	49	   ?		0
> 4	20070611	        A	50	   ?		0
> 
> (better picture here http://gyazo.com/5dc1218534d8b185ab22989069034b8a.png )
> 
> i need to fill "prior" column with following:
> - amount of hours spent on same activity by same user(ID) prior to
> date of current activity.
> 
> example:
> in case of observation 4(line 4) activity is B, so prior = sum of
> hours spent on activity B before it took place (i.e. before 20071210)
> there are 2 such observations, one with 49 hour, another with 24.
> Therefore, prior = 24+49=73.
> 
> 
> any suggestions how to make this simple?
> 
> i was thinking to use "sort" and "by:", but i can't figure out how to
> conditionally("if")  sum up values from different observations into
> "prior" var
> its not a first time i encounter similar problem and always took some
> way around, but this time i decided to ask, maybe there is proper,
> simple way to do this
> 
> thank you in advance
> 
> K.
> *
> *   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/