Hi Billy,
Ferret does not read dates such as "05-Sep-2007". The months are
numeric for these formats - I've added a note in the "set data"
documentation on that detail.
The separators should be able to be dashes or slashes, however
slashes seem to be what works. I will look into that and at least
update the documentation.
I converted your example to this file:
#sec central_date Miss ID corr Pmax Q(0-500m)
Q(0-700m)
1 05/09/2007 078006a 1 500 18.57
25.19
2 06/12/2007 07B018a 1 600 5.78
9.63
3 15/01/2008 07B018b 1 600 3.80
9.62
4 08/04/2008 082001a 1 600 -3.71
4.87
5 16/06/2008 082001b 1 600 16.43
16.63
6 21/07/2008 087006a 1 700 28.64
35.86
7 16/09/2008 087006b 1 700 14.80
15.31
8 01/12/2008 08B018a 1 700 16.57
22.62
9 23/01/2009 08B018b 1 700 4.22
6.70
10 02/08/2009 097006a 1 700 25.98
28.79
11 03/08/2009 097001a 1 700 24.83
28.62
And then it can be read with this command. These commands work with
Ferret back to v6.72; I think they should work with the versions you
have.
By the way, the name of this type is "EURODATE", for just a date
23/01/2009 or "EDATIME" for a combined euro-date-and-time (available
in v7.1 and higher), such as 22/01/2014T01:00:00. The documentation
has an example where it names a variable edate,
but its data type is eurodate.
yes? file/form=delim/delim="
"/skip=1/type="num,eurodate,text,num,num,num,num" dates.dat
yes? show data
currently SET data sets:
1> ./dates.dat (default)
name title I
J K L
V1 V1 1:20480
... ... ...
V2 V2 1:20480
... ... ...
(Julian days since 1-Jan-1900)
V3 V3 1:20480
... ... ...
V4 V4 1:20480
... ... ...
V5 V5 1:20480
... ... ...
V6 V6 1:20480
... ... ...
V7 V7 1:20480
... ... ...
yes? list v2
VARIABLE : V2 (days)
Julian days since 1-Jan-1900
FILENAME : dates.dat
SUBSET : 11 points (X)
1 / 1: 39328.
2 / 2: 39420.
3 / 3: 39460.
4 / 4: 39544.
5 / 5: 39613.
6 / 6: 39648.
7 / 7: 39705.
8 / 8: 39781.
9 / 9: 39834.
10 / 10: 40025.
11 / 11: 40026.
yes? define axis/t/t0=1-jan-1900/units=days taxis = v2
yes? show axis taxis
name axis # pts start
end
TAXIS TIME 11 i 05-SEP-2007 00:00
03-AUG-2009 00:00
T0 = 1-JAN-1900
Axis span (to cell edges) = 744.5
On 3/19/2018 4:07 PM, William S.
Kessler wrote:
I'm trying to read a file consisting of a list of values (first few lines listed below). I've tried with both Ferret 6.94 and PyFerret 7.1 (on a Mac). I want to turn the list into a netcdf file (on a time axis) that I can do subsequent work on.
The hard part is the formatted dates; I am unable to do this with SET DAT/FORMAT=DELIMITED
First there's the problem with the alphabetical months, but even after laboriously changing all of those to numerical, and the separating dashes to slashes, EDATE failed.
I finally kluged this by editing the text to enforce comma-separated values, but this seems like unnecessary work. This is especially true since after having read the edited file on the default X-axis, using the numerical day/month/year to define a time axis, RESHAPEing all variables to that, then writing as cdf, Ferret's listing of the first few values is remarkably like the original!
Namely, I did a bunch of manual editing to make a file that Ferret could read and convert back to much like what I had to start with.
Was there a more straightforward way to do this? Can Ferret make sense of ASCII dates like "05-Sep-2007"? Should have asked the list before I started?
Billy K
PS - sorry for my mailtool ruining the nice columns in both the original file and the Ferret listing.
-------------------
Original file (first 11 of hundreds of lines):
#sec central_date Miss ID corr Pmax Q(0-500m) Q(0-700m)
1 05-Sep-2007 078006a 1 500 18.57 25.19
2 06-Dec-2007 07B018a 1 600 5.78 9.63
3 15-Jan-2008 07B018b 1 600 3.80 9.62
4 08-Apr-2008 082001a 1 600 -3.71 4.87
5 16-Jun-2008 082001b 1 600 16.43 16.63
6 21-Jul-2008 087006a 1 700 28.64 35.86
7 16-Sep-2008 087006b 1 700 14.80 15.31
8 01-Dec-2008 08B018a 1 700 16.57 22.62
9 23-Jan-2009 08B018b 1 700 4.22 6.70
10 02-Aug-2009 097006a 1 700 25.98 28.79
11 03-Aug-2009 097001a 1 700 24.83 28.62
Ferret listing of resulting cdf:
yes? list/l=1:11 secnumt,midt,bott,t500t,t700t
DATA SET: ./solsea103-tr_v2b.cdf
TIME: 21-JUL-2007 00:00 to 27-AUG-2009 00:00
Column 1: SECNUMT is RESHAPE(SECNUM,TTHH)
Column 2: MIDT is RESHAPE(MID,TTHH)
Column 3: BOTT is RESHAPE(BOT,TTHH)
Column 4: T500T is RESHAPE(T500,TTHH)
Column 5: T700T is RESHAPE(T700,TTHH)
SECNUMT MIDT BOTT T500T T700T
05-SEP-2007 00 / 1: 1.00 "078006a" 500.0 18.57 25.19
06-DEC-2007 00 / 2: 2.00 "07B018a" 600.0 5.78 9.63
15-JAN-2008 00 / 3: 3.00 "07B018b" 600.0 3.80 9.62
08-APR-2008 00 / 4: 4.00 "082001a" 600.0 -3.71 4.87
16-JUN-2008 00 / 5: 5.00 "082001b" 600.0 16.43 16.63
21-JUL-2008 00 / 6: 6.00 "087006a" 700.0 28.64 35.86
16-SEP-2008 00 / 7: 7.00 "087006b" 700.0 14.80 15.31
01-DEC-2008 00 / 8: 8.00 "08B018a" 700.0 16.57 22.62
23-JAN-2009 00 / 9: 9.00 "08B018b" 700.0 4.22 6.70
02-AUG-2009 00 / 10: 10.00 "097006a" 700.0 25.98 28.79
03-AUG-2009 00 / 11: 11.00 "097001a" 700.0 24.83 28.62
|