On Sat, 12 Apr 2014, Logan Kelly wrote:
I have a quick question about haw gretl handles data imported from
the
BLS. If you download a series from the BLS, say into ms excel, you will
get years with 5 quarters or 13 months. The Q5, or M13 for monthly data,
is some annual summary of the data. What is the expected behavior of
gretl in the case that I import the data without properly removing the
non-observations, i.e. Q5/M13, from the csv file before importing? When
I try this experiment in the import output I find
"2001q05": BLS-type nonsense? Trying again
Does gretl drop these rows from the data set? That is what appears to be
happening. Also, is it terribly bad form to let gretl handle this issue
rather than removing the rows before import?
gretl grumbles, but it handles such data correctly (drops the offending
rows). If you don't mind the snark there's no need to pre-process BLS
data.
Allin Cottrell