[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: RE: RE: RE: Inputting arbitrary text files into Stata datasets
> Nick Cox, in response to Roger Newson (various extracts)
> > I think quotes are easier than you think. Compound double
> > quotes don't do any harm beyond adding some visual complexity
> > to what you read. If this is not true for you, there's a bug
> > somewhere.
> This is a problem. Consider:
> . local x `
> . di "`x'"
> . di `"`x'"'
> . di "`"
> . di `"`"'
This reminds of an old one. If you owe the bank $1000,
it's your problem, but if you owe the bank $1000 million,
it's the bank's problem.
In short, a problem for whom?
Here's what I think is Stata's view on this, and I have
to surmise in the absence of complete candour on its part.
Stata tells us its rules, including
1. ` ' delimit local macro names, " " ordinary strings
and `" "' any string.
2. Delimiters occur in pairs.
On the other hand, Stata is more than a list of explicit
rules, given that (for example)
a. The rules interact, occasionally in surprising ways.
b. There are (imperfect, undocumented) traps to trap some
usages frowned upon, as other threads indicate.
You broke the rules here. You didn't comment line
by line, but I count two of four cases which at first glance are
intuitive and two which are not. But that's a matter
of my intuition, which is of no concern to Stata.
In short, when you break the rules, Stata strains to understand
what you mean and sometimes fails. That's not what I call
a bug, for example.
* For searches and help try: