/*
	Biostat 212: Week 4 Data cleaning part 2
	Create Analysis Dataset do-file
	1) Run Cleaning do-file on pt1.dta and pt2.dta
	2) Append pt1.dta and pt2.dta
	3) Merge full pt dataset with doctors dataset 
	
*/
clear
capture log close
cd "/Users/crystal/Box/_Biostat212_2021/Week4/Lecture"
log using "output/week4lecture.smcl", replace

/*
	Open pt1.dta and clean dataset
*/

use "data/pt1.dta", clear
quietly run "do/cleaning.do" //runs cleaning.do file on pt1.dta data
save "data/pt1_clean.dta", replace

/*
	Open pt2.dta and clean dataset
*/
use "data/pt2.dta", clear
quietly run "do/cleaning.do" //runs cleaning.do file on pt2.dta data
save "data/pt2_clean.dta", replace

/*
	Open pt1_clean.dta
	Append pt2_clean.dta
*/
use "data/pt1_clean.dta", clear
append using "data/pt2_clean.dta"

/*
	Check appended datasets
*/
describe
browse

*Any variable that is not in both datasets will have missing values
tab nmorphine, missing
tab dis_date, missing

save "data/pts_complete.dta", replace

***************************************************************
/*
	Merge with doctor data using the "docid" variable
*/
***************************************************************
/*NOTE: class demo L4 Part 4 starts here. make sure you run everything 
above here BEFORE you start following along*/

*Check to see if we have multiple docid in our patient data
isid docid 
*OR*
duplicates report docid

*Have multiple docids so need to use m:1 (many to 1)
merge m:1 docid using "data/doctors.dta"
	tab _merge

*Drop if observation only was in the "doctors" dataset
drop if _merge==2

/*
	Save final dataset
*/

save "data/ptsanddocs_final.dta", replace
log close


//////////////////////////////////////////////

/*	
	Reshape the reshape_example.dta dataset from long to wide
*/

* Open the reshape example dataset
	use data/reshape_example.dta, clear 
		/*note: you need to remove quotes from the command when giving filepath. the following will 
		give you an error: 
			use "data/reshape_example.dta", clear 
		*/

* Explore dataset to verify it's in "long" format and what variables to use 
* for reshape command
	browse
	
* Reshape from long to wide
	reshape wide sbp, i(ptid) j(visit)

	browse // check to make sure this worked