The recent update of Stata included the new command -datasignature- as
advertised by William Gould. It has already saved me from one
potentially serious blunder. I was cleaning up a small survey data set,
correcting typos and checking for obvious errors and I saved the data
file and copied it to my memory stick. The next day, as I produced
histograms, I found another significant, but subtle, input error and I
corrected it and saved the data set (with the same name) but forgot to
copy it to my memory stick. When I went to work with the data at home,
I was using the old data but fortunately I had tried out -datasignature-
and I noticed that the numbers didn't match, since at the end of the
previous session I had copied the number down. My first reaction was
that -datasignature- didn't work, but then I realized my mistake.
Without it, I would have done my analysis with the wrong data set and
probably reported incorrect results. No doubt that I can be criticized
for not taking more care with my file names and data management, things
that I chide my students about. However, I now include -datasignature-
after each use and save.
This leads me to ask myself what other commands am I missing in my do
file boilerplate that I grab every time I start a new do file?
Currently I have:
// Project name and key word description
// data set to use
// date
// version 9.2
Capture log close
Set more off
Cd "copy from my documents"
// log using "name", text replace
Use data, clear
Datasignature
Summarize
In part it is a matter of taste, but am I missing something obvious or
important?
Douglas Dalenberg
Department of Economics
University of Montana
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
