Re: st: another coding que to group vars

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

It's probably easier to work with these data in long (rather than wide) format, but here's one way to do it:

clear


************************!
clear
inp ID  str1( Y1 Y2 Y3 Y4 Y5 Y6)  
1    x   x    x   x    x  x     
2   y    y  y    y    y  y      
3   x    x  x    x    .  x      
4  x    y   y   x    x  x       
5  y    y   y   x   x  x        
6  x    .   y    .    .  x      
7  x   y   x   x    .   .       
8  x  x   x   x    .    .       
9 y  y   y   y     .    .       
end

foreach v of varlist Y* {
	replace `v' = "0" if `v' == "y"
	replace `v' = "1" if `v' == "x"
}


destring, replace
g type = .
egen x = rowtotal(Y1-Y6)
egen y = rownonmiss(Y1-Y6)
replace y = y - x
egen first = rowfirst(Y1-Y6)


replace type = 1 if x > 0 & !y 
replace type = 2 if !x & y>0   
replace type = 3 if x>0 & y>0 & first
replace type = 4 if x>0 & y>0 & !first
replace type = 5 if x > 0 & !y & mi(Y5) & mi(Y6)
replace type = 6 if !x & y>0 & mi(Y5) & mi(Y6)
************************!


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


On Apr 11, 2012, at 5:50 PM, Rituparna Basu wrote:

> Thanks Eric! It looks like the 'group' command is similar to 'concat'.
> I used 'group' but it generated more than 1000 groups.
> My aim is to create the 'TYPE' variable such that  X  & Y are grouped into six broad categories,.  (type 1 is group X thru out(no matter the missing values), type 2 is group Y thru out (no matter the missing values), type 3 is started as group X and then changed to Y at any time point (no matter the missing values), type 4 is started as Y and then changed to X at any time point (no matter the missing values), type 5 & 6 have either X or Y thru out (no matter the missing values) except the last TWO years)).
> 
> 
> 
> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Eric Booth
> Sent: Wednesday, April 11, 2012 3:29 PM
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: another coding que to group vars
> Importance: High
> 
> 
> <>
> 
> Take a look at -egen- function group()
> 
> e.g.,  egen TYPE = group(Y1-Y6)  
> 
> - Eric
> 
> __
> Eric A. Booth
> Public Policy Research Institute
> Texas A&M University
> ebooth@ppri.tamu.edu
> Office: +979.845.6754
> 
> 
> On Apr 11, 2012, at 2:12 PM, Rituparna Basu wrote:
> 
>> Dear Statalist Scholars,
>> 
>> I have yet another coding question to ask. Please see the mock data below:
>> 
>> ID  Y1 Y2 Y3 Y4 Y5 Y6    TYPE     
>> 1    x   x    x   x    x  x      type1
>> 2   y    y  y    y    y  y       type2
>> 3   x    x  x    x    .  x        type1
>> 4  x    y   y   x    x  x        type3
>> 5  y    y   y   x   x  x         type4
>> 6  x    .   y    .    .  x          type3
>> 7  x   y   x   x    .   .           type3
>> 8  x  x   x   x    .    .           type5
>> 9 y  y   y   y     .    .           type6
>> 
>> 
>> Here ID= Patient ID
>> Y1-Y6= years
>> X & Y are two types of group
>> TYPE= This is the variable I would like to CREATE by grouping x & y into six categories.  (type 1 is group X thru out, type 2 is group Y thru out, type 3 is started as group X and then changed to Y at any time point, type 4 is started as Y and then changed to X at any time point, type 5 & 6 have either X or Y thru out except the last TWO years)). The types includes missing.
>> 
>> Any help is very much appreciated! Please let me know if more clarification is needed.
>> 
>> Thank you in advance!
>> 
>> Regards,
>> 
>> RB


*
*   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/