On Thu, 24 Sep 2009, Ofer Cornfeld wrote:
I have noticed that when importing EViews file the sample range is
set
incorrectly when first quarter in the year is the third.
I have attached an EViews 6 file ranging 1980Q3 - 1982Q3.
An import will set the Gretl range to 1980:1 - 1982:1 incorrectly.
Thanks. Unfortunately, getting date information out of Eviews
files is a guessing game, because they use an unpublished
proprietary binary format (which we have attempted to
reverse-engineer). I thought we had guessed how to find the
starting period correctly, but it seems not.
If you have access to Eviews, it might be helpful if you could
send me a few variant wf1 files: for example, a year's worth of
monthly data, in one case with the starting date set to 1990:01,
then with the start set to 1990:02, and so on. Hopefully this
would enable us to track the bytes where the starting period is
recorded.
Allin Cottrell