I wanted to change the sample by set range. I was not allowed to. I had to restructure the dataset. When the sample started at 1, I could select a subsample.
Gretl is getting confused by the fact that the observation numbers start at 5245. However, in principle that shouldn't be a problem. The confusion is now fixed in git, and new snapshots will follow before long.
BTW you could convert this dataset so that its 5-day daily structure is recognized via the following little script. The --omit-obs option to the store command ensures that the dates appear in the first column of tsx.csv, where they will be understood by gretl's open command as giving time-series information.
open TSX.gdt
store tsx.csv --omit-obs
open tsx.csv
store TSX_fixed.gdt
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I understand what you want, but that's exactly what gretl does by
default. If you do what I suggested you'll find that no missing values
are inserted. Gretl will report "Daily (5 days, incomplete)" but that
just means "incomplete" relative to a full 5-days-per-week calendar,
which is common with financial market data. We do offer the option of
inserting missing values, via
dataset pad-daily 5
but we assume that most of the time that's not what users will want.
I'm attaching a time-series plot; note that no missing values (gaps in
the lines) are to be seen.
I do not want to have missing values in the time series.
The sample is continuous for trading days.
I do not want to have gaps for days when the Canadian market was
not open.
Gretl is getting confused by the fact that the observation numbers start at 5245. However, in principle that shouldn't be a problem. The confusion is now fixed in git, and new snapshots will follow before long.
BTW you could convert this dataset so that its 5-day daily structure is recognized via the following little script. The --omit-obs option to the
storecommand ensures that the dates appear in the first column of tsx.csv, where they will be understood by gretl'sopencommand as giving time-series information.Hi Gábor,
I understand what you want, but that's exactly what gretl does by
default. If you do what I suggested you'll find that no missing values
are inserted. Gretl will report "Daily (5 days, incomplete)" but that
just means "incomplete" relative to a full 5-days-per-week calendar,
which is common with financial market data. We do offer the option of
inserting missing values, via
dataset pad-daily 5
but we assume that most of the time that's not what users will want.
I'm attaching a time-series plot; note that no missing values (gaps in
the lines) are to be seen.
Allin
On Sat, Aug 9, 2025 at 1:58 PM Gabor Korosi korosi.g@t-online.hu wrote: