pwd#
Session 1 videos:
video1 03/30/2023 (on courseplus)
video2 03/30/2023 (on courseplus)
profile.do Optional
video 04/20/2023 Includes exercise in debugging & use of `c(pwd)’
Some basic commands that the folks at Stata think you ought to know, by the end of this term. So keep checking on these every week to assess your progress.
Our goal in this session is to establish rules for all .do file scripts for the next 8 weeks:
StructureIndentationAnnotation
Every .do file will have these items (note: follow debugging guidelines when you get an error at if 1 {; watch 04/20/2023 video):
quietly {
qui {
}
if 0 {
qui {
if 0 { //background: documentation, not code
1. a few remarks
2. on what motivates
3. this .do file script
4. using ordinary text
5. stata will not read this block
}
}
if 1 {
qui {
if 0 { //background: context, not code
1. a few remarks
2. on what motivates
3. this .do file script
4. using ordinary text
5. stata will not read this block
}
if 1 { //methods: macros, logfile, settings
global workdir `c(pwd)' //change if different
capture log close
log using ${workdir}/pwd-demo.log,replace
set mem off
version 15
}
log close
}
if 2 {
qui {
if 0 { //background: context, not code
1. a few remarks
2. on what motivates
3. this .do file script
4. using ordinary text
5. stata will not read this block
}
if 1 { //methods: macros, logfile, settings
global workdir `c(pwd)' //change if different
capture log close
log using ${workdir}/pwd-demo.log,replace
set mem off
version 15
}
if 2 { //results: data, shape, etc.
use ${workdir}/transplants.dta, replace
di "shape: `c(N)' rows x `c(k)' columns"
}
log close
}
The if 1 { block is where you’ll typically define global macros, open a logfile, and determine either your pwd, filepaths, or urls
It will lay out the method or strategy by which you approach the task outlined in if 0 {
Each numerically sequenced if statement followed by a { will serve as your lodestar as you build a vertically growing script as contrasted with a horizontally sprawling one. Within each if int { you’ll have a block of code that serves a rather specific function and can be copied and pasted to another .do file when such a task is needed in the future or by someone else.
As you put your .do file together you’ll indent (typically 4 spaces) at each level: from qui { to if 0 {, and then from each if int { to the block of code within.
Within any block of code you’ll also indent as lower-level options and suboptions emerge in your script.
Look to the scripts throughout this book as exemplars. Notice that each block of code must be lightly annotated to describe what it does. Use the if 0 { for more detailed annotation.
Because this is an academic course, we will nudge you towards these practices by removing points in your homework if you neglect these three things: structure, indentation, and annotation.
qui {
if 0 { //background: documentation, not code
1. set the context of the entire .dofile
2. not the traditional way its done is Stata
3. this is mere mimicry of fancier notebooks
4. particularly Jupyter notebooks used for Python
5. as well as RStudio used for R
6. our dear Stata has no equivalent notebook
7. in such notebooks text-blocks & code-blocks
8. might be interspersed: a much richer way to
9. comment/annotate than you might be used to
}
if 1 { //methods: macros, logfile, settings
global workdir `c(pwd)' //change if different
capture log close
log using ${workdir}/pwd.log,replace
set mem off
version 15
pwd //informative, not usable
di c(pwd) //usable as macro, see above
ls -l
clear
cat adult.do //derived from program in chapter: delimit
do adult.do
di c(N) " rows & " c(k) " columns"
local shape: di c(N) " rows & " c(k) " columns"
di "`shape'"
}
if 2 { //continuous vars
lookfor age
local v: var lab HSAGEIR
di "`v'"
sum HSAGEIR, detail
return list
di "`r(p25)'"
putexcel set table1, replace
putexcel A1="Characteristics"
putexcel A2="`v'"
putexcel B1="(N=`c(N)')"
putexcel B2="`r(p50)' (`r(p25)'-`r(p50)')"
ls -ahl
}
if 3 { //binary vars
lookfor sex
codebook HSSEX
replace HSSEX=2-HSSEX
sum HSSEX
return list
local v: var lab HSSEX
di "`v'"
local per: di %3.1f r(mean)*100
di "`per'"
putexcel A3="`v'"
putexcel B3="`per'"
}
if 4 { //multicat vars
lookfor race
desc DMARETHN
capture label drop race
#delimit ;
lab def race
1 "White"
2 "Black"
3 "Hispanic"
4 "Asian";
lab values DMARETHN race;
#delimit cr
local v: var lab DMARETHN
putexcel A4="`v', %"
local vl: value label DMARETHN
levelsof DMARETHN, local(race)
return list
count if !missing(DMARETHN)
global N=r(N)
local row=5
foreach l of numlist `race' {
local per: lab `vl' `l'
qui sum DMARETHN if DMARETHN==`l'
local percent: di %3.1f r(N)*100/$N
di "`per', % : `percent'"
putexcel A`row' = " `per'"
putexcel B`row' = "`percent'"
local row=`row' + 1
}
}
if 0 { //blocks 2-4
1. all the code snippets you need
2. to confidently answer hw1.q5
3. revisit these now that you
4. know a little more Stata than
5. you did in wk1 when we run this script
}
log close
}