Re: [Gretl-users] csv date parsing
A somewhat belated response, but...
On Wed, 19 Nov 2008, Sven Schreiber wrote:
right now I am struggling a little bit with importing daily time
series data from a csv file. First of all it turns out that the
necessary format for dates is not explained in the gretl guide
for daily data.
After a lot of trial and error I just tried to imitate what
gretl itself writes out when exporting the example djclose.gdt
dataset, namely "2008/10/13", including the enclosing quotes.
That was actually a bit tricky, but with the rectangular
selection features of jEdit I eventually managed. (Yes jack, I'm
sure it's also possible with emacs :-)
Even after naming the first column 'obs', still it doesn't work!
Could it be that the problem is that my data is stored in
reverse chronological order (t=T at the top, t=1 at the bottom)?
Yes, that is (or was?) the entire problem right there. gretl has,
I think, pretty good heuristics for handling dates in a variety of
formats (e.g. YYYYMMDD, MMDDYY[YY], DDMMYY[YY], with or without
quotes, and using "/", ":" or "-" as separator, and also a
variety
of ways of expressing <year><quarter> or <year><month>).
It's just that up till now we've not attempted to read data
presented backwards.
Allin.