My point was not that you were advocating overwriting but that it might
be thought that I was...
And yes, I agree: Simplicity can be in the eye of the beholder.
I was thinking largely of brevity, you, it seems, largely of the
familiarity of commands you know well. But -split- might be a strange
command to some.
On -hexdump-: I don't think it's quite as difficult as you imply. For
example, with "@" as a candidate, you just need to use the -results-
option and check e.g. that r(c64) == 0.
All amicably, of course,
Nick
[email protected]
Joseph Coveney
Nick Cox wrote:
I would never recommend alteration of the original data file. I would
always recommend that you work on a copy. With that proviso, this route
is more complicated than using -filefilter-.
------------------------------------------------------------------------
--------
Lest there be a misunderstanding: I wasn't advocating overwriting an
original data file (or even a working copy) with -filefilter-. I don't
know
about analogous Unix commands, but have to assume that they, too, have
infile=, outfile= among their parameter lists.
We'll have to agree to disagree about relative complexity. Amicably, I
hope. I just find stopping to manually parse through -hexdump ,
tabulate-
output hoping that I don't overlook successive members of a list of
candidate substitutes, and then setting up -filefilter- with the winner,
and
then -insheet-ing the outfile, to be a similarly complicated prospect.
Maybe it's the inherent error-proneness of manually processing parallel
lists looking for things that aren't supposed to be there in one of them
that bothered me.
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/