Re: [ferret_users] Plotting Times

[Ansley Manke <ansley.b.manke@xxxxxxxx>]

Hi -
Short of parsing the strings within Ferret to get numerical dates and times (not something I'd want to do, myself!), you'd need to change the file a bit to get the delimited read to input this.  There's a stray quotation mark after the PM times, by the way.

The /FORMAT=date  does not translate month strings, only numerical month values. And the /TYPE=time does not translate AM and PM; so you'll need to handle that yourself.

So if I change the file to look like this:

09/29/2010,7:02 AM,6:54 PM,0.00,68,55,61,59,29.96
09/30/2010,7:03 AM,6:52 PM,3.40,78,64,71,76,29.52
10/1/2010,7:04 AM,6:51 PM,0.17,72,56,64,61,29.72
10/2/2010,7:05 AM,6:49 PM,0.00,73,49,61,56,29.99
10/3/2010,7:06 AM,6:48 PM,0.46,65,49,57,73,30.01
10/4/2010,7:07 AM,6:46 PM,0.45,55,48,52,82,30.10
10/5/2010,7:08 AM,6:44 PM,0.00,61,48,55,63,30.90
10/6/2010,7:09 AM,6:43 PM,0.05,60,51,56,67,29.95
10/7/2010,7:10 AM,6:41 PM,0.01,75,50,63,63,29.56
10/8/2010,7:11 AM,6:40 PM,0.00,80,48,64,56,30.02
10/9/2010,7:12 AM,6:38 PM,0.00,82,50,66,56,29.97
10/10/2010,7:13 AM,6:37 PM,0.00,81,50,66,63,29.93
10/11/2010,7:13 AM,6:35 PM,0.00,85,56,71,66,29.84
10/12/2010,7:14 AM,6:34 PM,0.00,82,58,70,78,29.75
10/13/2010,7:15 AM,6:32 PM,0.00,78,49,64,64,29.90
10/14/2010,7:16 AM,6:31 PM,0.96,57,48,53,92,29.87
10/15/2010,7:17 AM,6:29 PM,0.00,64,48,56,64,29.87
10/16/2010,7:18 AM,6:28 PM,0.00,67,47,57,58,29.92
10/17/2010,7:19 AM,6:27 PM,0.00,79,44,62,69,29.94
10/18/2010,7:21 AM,6:25 PM,0.00,68,47,58,71,29.98
10/19/2010,7:22 AM,6:24 PM,0.00,61,51,56,82,29.62

Then letting both commas and spaces be the delimiters, and putting in a dummy variable for the letters AM, here are lines that read just the date and the two times.

yes? file/form=delim/delim="\b,\,"/var="date,timeam,dum,timepm"/type="date,time,text,time"   input.dat

yes? ! Define a time axis from date, and put the times onto that axis -
yes? define axis/t0=1-jan-1900/units=days taxis = date

yes? let/title=hours_AM time_am = reshape(timeam, t[gt=taxis])
yes? let/title=hours_PM time_pm = 12+ reshape(timepm, t[gt=taxis])
yes? plot/vlim=0:24 time_am, time_pm

Or do a PLOT/VS or whatever you'd like from there

-Ansley

On 10/26/2010 12:12 PM, Mark Verschell wrote:

Hi Ansley!

	No, it's an ASCII file. I tried it several ways.  Here is a brief section from my latest attempts. I was trying to plot V2 vs time. I realize this particular example won't work, since I am trying to plot strings, but I couldn't come up with a way to basically plot time vs date (and I was getting desperate)! I eventually did this in a spreadsheet (where I could do a scatter plot, but not a line plot!) But I still would like to know how to do this in ferret.

"Sep 29 2010","7:02 AM",6:54 PM",0.00,68,55,61,59,29.96
"Sep 30 2010","7:03 AM",6:52 PM",3.40,78,64,71,76,29.52
"Oct 1 2010","7:04 AM",6:51 PM",0.17,72,56,64,61,29.72
"Oct 2 2010","7:05 AM",6:49 PM",0.00,73,49,61,56,29.99
"Oct 3 2010","7:06 AM",6:48 PM",0.46,65,49,57,73,30.01
"Oct 4 2010","7:07 AM",6:46 PM",0.45,55,48,52,82,30.10
"Oct 5 2010","7:08 AM",6:44 PM",0.00,61,48,55,63,30.90
"Oct 6 2010","7:09 AM",6:43 PM",0.05,60,51,56,67,29.95
"Oct 7 2010","7:10 AM",6:41 PM",0.01,75,50,63,63,29.56
"Oct 8 2010","7:11 AM",6:40 PM",0.00,80,48,64,56,30.02
"Oct 9 2010","7:12 AM",6:38 PM",0.00,82,50,66,56,29.97
"Oct 10 2010","7:13 AM",6:37 PM",0.00,81,50,66,63,29.93
"Oct 11 2010","7:13 AM",6:35 PM",0.00,85,56,71,66,29.84
"Oct 12 2010","7:14 AM",6:34 PM",0.00,82,58,70,78,29.75
"Oct 13 2010","7:15 AM",6:32 PM",0.00,78,49,64,64,29.90
"Oct 14 2010","7:16 AM",6:31 PM",0.96,57,48,53,92,29.87
"Oct 15 2010","7:17 AM",6:29 PM",0.00,64,48,56,64,29.87
"Oct 16 2010","7:18 AM",6:28 PM",0.00,67,47,57,58,29.92
"Oct 17 2010","7:19 AM",6:27 PM",0.00,79,44,62,69,29.94
"Oct 18 2010","7:21 AM",6:25 PM",0.00,68,47,58,71,29.98
"Oct 19 2010","7:22 AM",6:24 PM",0.00,61,51,56,82,29.62

BTW, I read it with FILE/FORMAT=DELIM

Mark Verschell
mav9@xxxxxxx

"What we have done for ourselves alone dies with us. What we have done for others and the world remains and is immortal." - Albert Pines



On Oct 26, 2010, at 12:49 PM, Ansley Manke wrote:

Hi Mark,
That sounds as if some of the data are strings? If it's in NetCDF, what does the ncdump header of the dataset look like?

Ansley

On 10/25/2010 5:44 PM, Mark Verschell wrote:

I have a data set that lists a time of day versus day of the year (i.e., what time of day something happened each day over a year).

What I am looking for is how to plot this time series. Does anyone have any ideas.

Further information, I tried this various ways, and I continuously get the error message:

"**ERROR: illegal data type (float,string,...) for operation: V2"

Mark Verschell
mav9@xxxxxx

"What we have done for ourselves alone dies with us. What we have done for others and the world remains and is immortal." - Albert Pines