Re: [ferret_users] strange interaction between viewport and TXTYPE

The code automatically decides whether to use the 3-letter month labels or the 1-letter ones, depending on how much space is available. For instance, if the script try2.jnl includes

then the style with JAN FEB MAR will consistently be used. The same happens for a default plot with time on the horizontal axis. Or you can use the PPL txtype command, as in the first post here, to set the time-axis label style and make sure the setting stays consistent.

It looks to me as if the y-axis length defined for the viewport somehow gets changed from the first running of the commands to later ones. Without the addition of SET WINDOW/ASPECT command, I see this,

And with subsequent calls to try2, that second set of lengths is shown. So the choice of 3-letter labels or 1-letter labels is being made depending on the setting of y-len at that time. I don't know why the axis length might be changing. But, with the "set aspect" command in there, the axis-lengths are consistent and the time-axis plotting is also consistent.

(The content of this message are mine personally. I am not affiliated with PMEL.)

PyFerret v7.63 (optimized)

Darwin 19.6.0 - 10/13/20

-> Same results as Ryo

FERRET v7.2 (optimized)

Linux 2.6.32-696.3.2.el6.x86_64 64-bit - 07/11/17

-> Get J, F, M, ... at each try

PyFerret v7.43 (optimized)

Linux 3.10.0-862.11.6.el7.x86_64 - 09/14/18

-> Same results as Ryo

FERRET v6.72

Linux 2.6.32-131.12.1.el6.x86_64 64-bit - 09/13/11

Linux 2.6.32-696.3.2.el6.x86_64 64-bit - 07/11/17

-> Get J, F, M, ... at each try

Le 23 nov. 2020 à 15:40, Ryo Furue <furue@xxxxxxxxxx> a écrit :

Dear Ferreters,

1) It seems that "set viewport" changes the type of the time-axis labels.

Without "set viewport", I get J, F, M, . . . . for my climatological axis; with set viewport, I get JAN, FEB, MAR, . . .

2) Run the same script again and get different time-axis labels.

When I run the same script again, I get J, F, M, . . . !!! I don't have anything in my ~/.ferret .

Here is my script. It's self-contained and doesn't need external datafiles.

! --- try2.jnl ---

set data coads_climatology
define view/axes/x=0.5:0.9/y=0.1:0.8 myright
set view myright
fill/transpose/hlimits=22S:34S SST[x=180]

and you run it as

yes? go try2.jnl ! -> JAN, FEB, MAR, . . .

yes? go try2.jnl ! -> J, F, M, . . .

To get a stable result, I seem to always need

PPL txtype,YR,MON1

I used

PyFerret v7.6 (optimized)
Darwin 18.7.0 - 06/25/20

on macOS 10.15.7

for this test.

Regards,

Ryo

—

Olivier Marti - LSCE Bât 714 p. 1049

MERMAID Team

Normal situation : +33 1 69 08 77 27

Corona lockdown : +33 6 45 36 43 74

mailto:olivier.marti@xxxxxxxxxxxx