[ncl-talk] Sequences of values

Kiswendsida Hyacinthe GUIGMA karongseba at gmail.com
Mon Mar 26 09:45:59 MDT 2018

```Hi Rashed and everyone,Thank you for your suggestion.I also had the idea of using many loops. It works perfectly for no large data. But when it comes to apply it to a 3D data covering half of the Africa domain, it slows down the execution and ends up killing the process. That's why I was thinking that there exist an other way to do that using not as much iterations.
Best

-----------------------------
GUIGMA

On Saturday, 24 March 2018, 18:30:38 GMT, Rashed Mahmood <rashidcomsis at gmail.com> wrote:

Hi Guigma,
For 1 dimensional series you can use the following. however, you would need to adapt this to your 3D data sets through loops/ind function and need to do TESTING, this is a very time consuming process.
begin
;  X1 = (/0,1,0,0,1,1,1,1,1,0,0,1,0,1,1,1,1,0,0,1,0/) ; orignal
X1 = (/0,1,0,0,1,1,1,1,1,0,0,1,0,1,1,1,1,0,0,1,0,1,1,0,0,1,0,1,0,1,1/) ; for testing

X2 = new(dimsizes(X1),integer)

do n=0,dimsizes(X1)-1
if(n.eq.0)
if(X1(n).eq.1.and.X1(n+1).eq.0)
X2(n) = 0
else
X2(n) = X1(n)
end if
end if

if(n.gt.0.and.n.lt.dimsizes(X1)-1)
if(X1(n).eq.1)
if(X1(n-1).eq.0.and.X1(n+1).eq.0)
X2(n) = 0
else
X2(n) = X1(n)
end if
else
X2(n) = X1(n)
end if

end if

if(n.eq.dimsizes(X1)-1)
if(X1(n).eq.1.and.X1(n-1).eq.1)
X2(n) = X1(n)
else
X2(n) = 0
end if
end if
end do
print(X1+" ... "+X2)

end

Note I have NOT tested it as much one should so I would let you to test and find out if that is what you are looking for.

Cheers,
Rashed

On Sat, Mar 24, 2018 at 6:44 AM, Kiswendsida Hyacinthe GUIGMA <karongseba at gmail.com> wrote:

Thank you for your reply and for your help. Actually, before to post my question, I tried to make a use of dim_numrun_n but it didn't solve my problem.
You misunderstood a bit my [1]. I said that I successfully did that... So it is OK for that.
For [2] , what the dim_numrun_n  function allows is just to do a count. What me I'd rather like is to set to _FillValue (or to 0 if use of where function) all values that are not in a sequence of at least 04 days. As an example I would like to have something as follows:X1=(/0,1,0,0,1,1,1,1,1,0,0,1, 0,1,1,1,1,0,0,1,0,/) ==========> X2= (/0,0,0,0,1,1,1,1,1,0,0,0,0,1, 1,1,1,0,0,0,0,/).In other words it is to replace all "isolated" 1s by 0.
Kindest regards

-----------------------------
GUIGMA

On Friday, 23 March 2018, 22:17:11 GMT, Dennis Shea <shea at ucar.edu> wrote:

Sorry, I'm I am not looking at any code.

tcrit = 30.

T = f->T                  ; (time,lat,lon)

[1] "setting values below the threshold to _FillValue" is simple.

https://www.ncl.ucar.edu/ Document/Functions/Built-in/ where.shtml

Tcrit = where(T.ge.tcrit, T, T at _FillValue)
copy_VarMeta(T, Tcrit)
===========
[2] You do not need to do [1]

Maybe you could use the following function:

=========
http://www.ncl.ucar.edu/ Document/Functions/Built-in/ dim_numrun_n.shtml
***Look at the examples***
=========

T10= where(T.ge.tcrit,1, 0)
T10 at long_name = "Exceed=1; No=0"
copy_VarCoords(T, T10)
printVarSummary(T10)
r0 = dim_numrun_n (T10, 0, 0)  ; (ntim,nlat,mlon)
r1 = dim_numrun_n (T10, 1, 0)

copy_VarCoords(T, r0)
copy_VarCoords(T, r1)

;---Pick one grid point to see what is happening

LAT = ..
LON = ..
print(T(:,{LAT},{LON})+"  "+Tcrit(:,{LAT},{LON})+"  "+T10(:,{LAT},{LON}) \
+"  "r0(:,{LAT},{LON})+"   "+r1(:,{LAT},{LON})

On Fri, Mar 23, 2018 at 3:29 PM, Kiswendsida Hyacinthe GUIGMA <karongseba at gmail.com> wrote:

Dear NCL users,I have 3D (time,lat,lon) array of temperature from which I would like to keep only sequences (consecutive days) exceeding a certain threshold.  This consist concretely in setting values below the threshold to _FillValue.  The length of sequence is set to 4. So far,  have been able  to set to _FillValu only individual days that exceed the threshold. Then I would be pleased if you could provide me with a method to also set to _FillValue values that are not into sequences of at least 04 days.
Please find attached my data and my code.

Regards

-----------------------------
GUIGMA
______________________________ _________________
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/20180326/a815e12c/attachment.html>
```