Hi David,
If a time axis is less than a year long, and if it starts in year 0 or
year 1, then it's treated as a Modulo time axis, e.g. a climatology. In
a climatology, the year is meaningless and so it's not listed.
Unfortunately this behavior is carried over into longer time axes.
I'll put this into our list of bugs to fix.
yes? define
axis/t0=1-jan-0000/t="15-jan-0000":"15-jan-1900":4/units=months long_tax
*** NOTE: /UNIT=MONTHS is ambiguous ... using 1/12 of 365.2425 days
! The first 6 points are in year 0000 and 0001, but the year isnt
listed.
yes? list/l=1:15 t[gt=long_tax]
VARIABLE : T
axis LONG_TAX
SUBSET : 15 points (TIME)
15-JAN / 1: 0.46
15-MAY / 2: 4.46
14-SEP / 3: 8.46
14-JAN / 4: 12.46
15-MAY / 5: 16.46
14-SEP / 6: 20.46
14-JAN-0002 / 7: 24.46
16-MAY-0002 / 8: 28.46
14-SEP-0002 / 9: 32.46
14-JAN-0003 / 10: 36.46
16-MAY-0003 / 11: 40.46
15-SEP-0003 / 12: 44.46
14-JAN-0004 / 13: 48.46
15-MAY-0004 / 14: 52.46
14-SEP-0004 / 15: 56.46
David Wang wrote:
Hi,
The starting year is actually 1, not 0. I was mistaken. But the problem
persists. I wonder why Ferret drops off 0001 in time axis strings.
yes? show g/l=1:24 temp
GRID GJG1
name axis # pts start end
LON LONGITUDE 1mr 0E 0E
LAT LATITUDE 1 r 0 0
DEPTH DEPTH (m) 1 r- 5 5
TIME TIME 6000 i 16-JAN 12:00 16-DEC-0500
12:00
L T TBOX TBOXLO TSTEP
(days)
1> 16-JAN 12:00:00 29.5 01-JAN 18:00:00
15.5
2> 15-FEB 00:00:00 29.5 31-JAN 06:00:00 45
3> 16-MAR 12:00:00 30 01-MAR 18:00:00
74.5
4> 16-APR 00:00:00 30.5 31-MAR 18:00:00
105
5> 16-MAY 12:00:00 30.5 01-MAY 06:00:00
135.5
6> 16-JUN 00:00:00 30.5 31-MAY 18:00:00
166
7> 16-JUL 12:00:00 30.75 01-JUL 06:00:00
196.5
8> 16-AUG 12:00:00 30.75 01-AUG 00:00:00
227.5
9> 16-SEP 00:00:00 30.5 31-AUG 18:00:00
258
10> 16-OCT 12:00:00 30.5 01-OCT 06:00:00
288.5
11> 16-NOV 00:00:00 30.5 31-OCT 18:00:00
319
12> 16-DEC 12:00:00 30.75 01-DEC 06:00:00
349.5
13> 16-JAN-0002 12:00:00 30.25 01-JAN-0002 00:00:00
380.5
14> 15-FEB-0002 00:00:00 29.5 31-JAN-0002 06:00:00
410
15> 16-MAR-0002 12:00:00 30 01-MAR-0002 18:00:00
439.5
16> 16-APR-0002 00:00:00 30.5 31-MAR-0002 18:00:00
470
17> 16-MAY-0002 12:00:00 30.5 01-MAY-0002 06:00:00
500.5
18> 16-JUN-0002 00:00:00 30.5 31-MAY-0002 18:00:00
531
19> 16-JUL-0002 12:00:00 30.75 01-JUL-0002 06:00:00
561.5
20> 16-AUG-0002 12:00:00 30.75 01-AUG-0002 00:00:00
592.5
21> 16-SEP-0002 00:00:00 30.5 31-AUG-0002 18:00:00
623
22> 16-OCT-0002 12:00:00 30.5 01-OCT-0002 06:00:00
653.5
23> 16-NOV-0002 00:00:00 30.5 31-OCT-0002 18:00:00
684
24> 16-DEC-0002 12:00:00 30.75 01-DEC-0002 06:00:00
714.5
D.
On Mon, Feb 25, 2008 at 1:18 PM, David Wang
< climater@xxxxxxxxx>
wrote:
Hello
Ferreters,
I'm using `var,return=tstart,p=1` to return the starting year and
likewise the ending year. However when the starting year is 0, the
command returns nothing (a null string) which causes troubles when I'd
like to assign the value to other variables/symbols. My intention is to
calculate multi-year means and variances from monthly data by the
approach described here at http://ferret.pmel.noaa.gov/FERRET_17sep07/FAQ/analysis/annual_from_monthly.html
I need the starting and the ending years of the original time axis to
define a new time axis. Any clues?
yes? say `temp,return=tstart`
!-> MESSAGE/CONTINUE 16-JAN 12:00
16-JAN 12:00
yes? say `temp,return=tstart,p=1`
!-> MESSAGE/CONTINUE
Thanks very much,
David
|