As its original parent, I am a little queasy
at the thought of adding yet another option
to -destring-, already a bit more complicated
than perhaps is desirable.
However, as Uli correctly states, this command
is now under the hand of StataCorp.
However, -destring- was always intended to be as safe as
possible.
Although this format is often called "European", it
is not used in all countries of Europe, so I have
some questions for those familiar with it to inform
those at StataCorp considering this change.
1. If the -comma- option is intended as a special option for
numbers in which the comma indicates the decimal
place, shouldn't the replacement take place at
most once? That is, any string with 2 or more
commas cannot be a valid number and should trigger
an error message.
2. Do other marks ever appear, e.g. stops (periods)?
Nick
[email protected]
Ulrich Kohler
> Seb Buechte wrote:
> > > Common problem for european users. The easier way to
> avoid it: using
> > > Excel, apply function "replace" comma by point (replace all) then
> > > save as text file (tab delimited). open stata and import your .txt
> > > file will work fine.
> >
> > Why is that easier?
>
> Among research groups this depends on _who_ has to do it ;-)
>
> Probably StataCorp might add a comma-option to -destring-, hence:
>
> -------------------------------------------destring.ado
> program destring
> syntax [varlist], [Generate(string) replace] [force] [float] /*
> */ [Ignore(string)] [percent] [comma]
>
> if "`comma'" ~= "" {
> foreach var of local varlist {
> replace `var' = subinstr(`var',",",".",.)
> }
> }
>
> <go on as in destring.ado>
> -----------------------------------------------------------
>
> As End User you might also copy destring.ado to
> [personal]/destring2.ado and
> make the changes there.
*
* 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/
