[ncl-talk] Loop Breaking on 3rd pass

[Rick Brownrigg]

I don't think that's quite right; you've got this:
    ....
    n_years = dimsizes(years)
    WRF_tot = new(n_years-1, "float")
    do i = 0, n_years-1
        s_date1 = years(i) + "110100"
        e_date1 = years(i+1) + "033100"
    ....

n_years is the size of the array years, but the valid indices into the
array are [0..n_years-1]. So on the last iteration of the loop, i=nyears-1,
and i+1 = n_years which is outside the range of valid indices. You should
be getting an error message like:

    fatal:Subscript out of range, error in subscript #0

Are you not?





On Thu, May 9, 2019 at 8:05 PM Zach Rieck <zrr817 at gmail.com> wrote:

> That was my thought initially as well, but I actually have the code wired
> to handle that. The loop only runs for the length of the years -1, so i+1
> should pull in the last value.
>
> I didn't include the error since it was vague and didn't seem to tell me
> anything. All it said was fatal exec and pointed to the line where I pull
> on the gridpoint value in the loop.
>
> -Zach Rieck
> zrr817 at gmail.com
> (513)-502-5652
>
> On Thu, May 9, 2019, 7:16 PM Rick Brownrigg <brownrig at ucar.edu> wrote:
>
>> Hi Zach,
>>
>> Its always helpful to include error messages.  If I were to speculate,
>> I'd say its probably this line:
>>
>>     e_date1 = years(i+1) + "033100"
>>
>> The "i+1" probably exceeds the dimensionality of the variable years
>>
>> Rick
>>
>>
>> On Thu, May 9, 2019 at 6:11 PM Zach Rieck <zrr817 at gmail.com> wrote:
>>
>>> That just about did it! I made a few minor changes to the array lengths
>>> and added the colon and it now runs until the final pass through before
>>> breaking. I'm not sure why though, I can query all the values for that pass
>>> without any issue.
>>>
>>> Attached the updated code with the changes. Thanks for the help and
>>> patience!
>>>
>>> -Zach Rieck
>>> zrr817 at gmail.com
>>> (513)-502-5652
>>>
>>>
>>> On Thu, May 9, 2019 at 7:51 AM Dennis Shea <shea at ucar.edu> wrote:
>>>
>>>> NCL is strongly typed. By default, NCL will not allow different size
>>>> arrays to the left and right of the assignment operator [ *=* ].
>>>>
>>>> Use NCL's reassignment operator   *:=*
>>>>
>>>> WRF_gridpoints *:=* f_in[:]->PREC_ACC_NC(s_time3:e_time3,553,412)
>>>> ===
>>>> *https://www.ncl.ucar.edu/Document/Language/varb_assign.shtml*
>>>> <https://www.ncl.ucar.edu/Document/Language/varb_assign.shtml>
>>>>
>>>> On Wed, May 8, 2019 at 7:02 PM Zach Rieck <zrr817 at gmail.com> wrote:
>>>>
>>>>> To Whom it May Concern-
>>>>>
>>>>> I have a loop where I am trying to pull precip for different time
>>>>> periods for each winter. The loop is successfully pulling in the write
>>>>> dates and totals, and runs EXACTLY how it is supposed to until the 3rd pass
>>>>> through. I'm confused by this since if the loop wasn't repeatable it should
>>>>> error out no later than the 2nd pass. The error I am getting is that on the
>>>>> 3rd pass, the left and right side dimensions do not match when I try to
>>>>> pull in the gridpoint data
>>>>>
>>>>> I have verified the lengths on my variable and all are long enough to
>>>>> pass through the loop as they are meant to. The loop is at the end of the
>>>>> code, from ~ line 73-78.
>>>>>
>>>>> Thanks for the help!
>>>>>
>>>>> -Zach Rieck
>>>>> zrr817 at gmail.com
>>>>> (513)-502-5652
>>>>> _______________________________________________
>>>>> ncl-talk mailing list
>>>>> ncl-talk at ucar.edu
>>>>> List instructions, subscriber options, unsubscribe:
>>>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>>>
>>>> _______________________________________________
>>> ncl-talk mailing list
>>> ncl-talk at ucar.edu
>>> List instructions, subscriber options, unsubscribe:
>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20190509/85093daa/attachment.html>