> Dear all,
>
> I have corporate announcements data where records of a string
> var are inconsistently capitalised, which makes analysis
> difficult. To illustrate, the following records should really
> appear as one and the same:
>
> Annual Report and Accounts
> Annual Reports and Accounts
> Annual report and accounts
> Annual reports and Accounts
> Annual reports and accounts
>
> I can globally replace 'reports' by 'report' in a text editor
> but would like to automate case conversion in Stata if I
> could - say, convert everything to lower case (or just the
> first word capitalised, or every word in the header).
> -renvars- does this but for variable names, not observations.
> -mixcase-, on the other hand, seems just the job, but it was
> written by Bill Gould nearly ten years ago (it was part of
> dm13.1) so I doubt it would run smoothly three or four
> versions of Stata later. Is anyone aware of an update to this
> routine or an alternative way of doing the same?
First, a note: in fact, Stata's version control generally ensures that
antique programs continue to function correctly. If you look at the
contents of mixcase.ado, you will see that the first line is "version
3.0"; so -mixcase- should work just fine. In my quick tests, it seems
to work OK for me.
But, there is an easier way -- the lower() function. That turns a
string into all lower case. So you could just:
replace <varname>=lower(<varname>)
If you have a lot of variables:
for var <varlist> : replace X = lower(X)
or
foreach v in <varlist> {
replace `v' = lower(`v')
}
That deals with the capitalization. For the "reports" and "report" and
the like,
I would use the subinstr() function. (I think that stands for
SUBstitute IN STRing):
replace <varname> = subinstr(<varname>,"reports","report",.)
replace <varname> = subinstr(<varname>,"reports","report",.)
These can be similarly wrapped up in the -for- or -foreach- loops.
--Nick Winter
*
* 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/
