EViews.com

hi,
I have a file that has only two daily datapoint for each series, but when I use command to open it, eviews automatically recognized it as monthly since the dates are
12/01/2012 and 11/01/2012.

I used the commands below to solve the problem but still the problem is there. Anyone knows how to let eviews to recognize it as daily?

Thanks a lot.

Commands used but not working:


(1) wfopen(d=page) "C:\path\commodity.d1.csv" colhead=3, namepos=last @id @date(@dateval(mnemonic,"DD/MM/yyyy"))

(2) wfopen(d=page) "C:\path\commodity.d1.csv" colhead=3, namepos=last @id @date(@dateval(mnemonic,"MM/DD/yyyy"))
(3) wfopen(d=page) "C:\path\commodity.d1.csv" colhead=3, namepos=last
pagestruct(freq=d5) @date(mnemonic)

I'm afraid the only way to fix things up is to manually tell EViews what the formats of the columns are so you can provide an explicit format for the date column. This should do the trick:

wfopen(d=page) "commodity.d1.csv" colhead=3,namepos=last,format=(1D.1,27W),datefmt="dd/mm/yyyy"

Once the date format is corrected, all the other automatic analysis should work fine.

The format description (1D.1, 27W) means:

- one date column "1D", following the first format in the date format list ".1" (without the ".1" EViews will try to automatically guess the date format)

- twenty-seven other columns "27W" , allowing EViews to automatically determine the formats (no good reason for the letter "W" - just that none of the more obvious letters were available)

Unfortunately you can't just override the format of the first column while leaving the rest of the format analysis on fully automatic mode,. This means that you will need to know how many columns are in the input file.

I agree this is kind of clunky - we should probably try to make this operation simpler somehow.