| Re: Log histogram = FAIL [message #72042 is a reply to message #71962] |
Tue, 03 August 2010 12:24   |
Haje Korth
Messages: 651
Registered: May 1997
|
Senior Member |
|
|
Chris,
just to confirm to the forum, that it is the histogram keyword that
causes the unwanted results. I still had this in memory from the tech
preview, but since it is not documented I cannot complain here. :-)
I will test the barplot routine instead. Hopefully one can turn off
the sides of the bars so my distributions do not look like cities with
sky scrapers.
Haje
On Aug 3, 2:17 pm, Chris Torrence <gorth...@gmail.com> wrote:
> On Aug 2, 12:17 pm, Haje Korth <hajeko...@gmail.com> wrote:
>
>> Plotting log histograms using new graphics system in IDL does not work
>> right. Try
>
>> IDL> a=findgen(10)+1
>> IDL> b=10./a^2
>> IDL> b[8]=0
>> IDL> p=plot(a,b,/histogram,/ylog,yrange=[1.0d-3,10])
>
>> Then tell me what happened to the last data point. I understand that 0
>> is an illegal log value, but IDL should be smart enough to filter
>> this. BTW, this works fine in direct graphics.
>
>> Haje
>
> Hi Haje,
>
> The log plot seems to work correctly without the histogram keyword:
>
> p=plot(a,b,'-o',/ylog,yrange=[1.0d-3,10])
>
> We didn't document the histogram keyword, because we wanted people to
> start using barplot() instead. Unfortunately, for IDL 8.0, BARPLOT
> doesn't handle logarithmic axes. However, I have just fixed the
> barplot code, and it will be available in the next IDL update.
>
> Cheers,
> Chris Torrence
> IDL Software Development Manager
> ITTVIS
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
