On Tue, 5 Mar 2013, Riccardo (Jack) Lucchetti wrote:
On Tue, 5 Mar 2013, Allin Cottrell wrote:
> On Tue, 5 Mar 2013, Andreas Noack Jensen wrote:
>
>> I am checking some gretl code against CATS in RATS and using the
>> australian
>> dataset. When exporting to csv, the variable iau2 only has three digits
>> after the dot, but the variable has up to four non zero digits after the
>> dot in gretl and hence I get different results from the exported data than
>> I do in gretl. Is it a bug?
>
> Looks like it. I'll take a look at the relevant code shortly.
In the meantime, you can use
set csv_digits 18
and force the number of digits in the csv file.
Hmm, yes, we need to document that.
Anyway, I found the bug. We were trying to be a bit too clever in
saving bytes in CSV output, and failed (occasionally) in quite a
subtle way. Fixed in CVS; snapshots will follow tomorrow.
Allin