Re: tick mark crash

Hi Mark,

The problem is due to one of your values being equal to infinity:

        ncl_filedump -v drake auscom_1-500_5_cnyf2-01_drake.nc

. . .
Variable No. 0
      drake:

    155.3 150.9 149.5 146.7 146.5 144.5
    143.2 141.6 140.2 139.6 138.8 138.8
    138.4 138.1 138.1 138.4 139.0 139.7
    140.3 140.9 141.5 142.3 142.8 143.6
    144.6 145.6 146.5 147.8 149.0 150.0
    151.1 152.1 153.2 154.1 154.9 155.7
    156.5 +Infinity 157.8 158.4 158.9 159.3
    159.7 160.2 160.7 161.2 161.6 162.0
    162.3 162.6 162.8 163.1 163.3 163.5
    163.7 163.7 163.8 164.0 164.2 164.4
    164.5 164.7 164.9 165.1 165.2 165.4
    165.5 165.6 165.7 165.8 165.9 166.0
    166.0 166.1 166.2 166.2 166.3 166.3
    166.4 166.4 166.4 166.5 166.5 166.5
    166.5 166.5 166.5 166.5 166.5 166.5
    166.4 166.4 166.5 166.4 166.4 166.3
    166.3 166.3 166.2 166.2

If I replace this value with auscom@_FillValue (auscom(37) =
auscom@_FillValue), then the script runs to completion.

Unfortunately, there's no way to check for an infinity value in NCL
(although you can test for Nan).

--Mary

On May 27, 2010, at 5:55 PM, mark collier wrote:

> Hi,
> I'm getting a crash with a particular dataset, however, another
> dataset very similar is accepted. I've attached a cut down script
> and small dataset. I'm using NCL5.2.0 on a sun constellation. I've
> tried earlier version of NCL and a MAC version with the same error.
> My attempts at changing tm resources didn't get anywhere. Any help
> appreciated
> Regards,
> Mark.
>
>
> <
> auscom_1
> -500_5_cnyf2
> -01_drake.nc><test.ncl>_______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Fri May 28 06:56:11 2010