NCL Home > Documentation > Functions > General applied math, Statistics

dim_cumsum

Calculates the cumulative sum along the rightmost dimension.

Prototype

	function dim_cumsum (
		x    : numeric,  
		opt  : integer   
	)

	return_val [dimsizes(x)] :  typeof(x)

Arguments

A variable of numeric type and any dimensionality.

opt

Option for how to treat _FillValue.

opt = 0 ; when _FillValue is encountered, set current element and all remaining to _FillValue
opt = 1 ; when _FillValue is encountered set element to _FillValue and resume summing
opt = 2 ; treat _FillValue as numeric zero when summing.

Return value

The same type and dimensionality as x.

Description

The dim_cumsum function calculates the cumulative sum of all elements of the n-1th (rightmost) dimension for each index of the dimensions 0...n-2. Missing values are treated differently as specified by opt.

Use dim_cumsum_n if you want to specify which dimension(s) to do the cumulative sum across.

Use the dim_cumsum_Wrap function if metadata retention is desired. The interface is identical.

Examples

Example 1: When no _FillValue are encountered all three options return the same result.

  x        = (/1,2,3,4,5/)
  x_sum    = sum(x)            ; result 15        
  x_cumsum = dim_cumsum(x, 0)  ; result (/1  3  6  10  15/)
  x_cumsum = dim_cumsum(x, 1)  ; result (/1  3  6  10  15/)
  x_cumsum = dim_cumsum(x, 2)  ; result (/1  3  6  10  15/)

  ; Use dim_cumsum_Wrap if metadata retention is desired
  ; x_cumsum = dim_cumsum_Wrap(x, 0)  ; result (/1  3  6  10  15/)
  ; x_cumsum = dim_cumsum_Wrap(x, 1)  ; result (/1  3  6  10  15/)
  ; x_cumsum = dim_cumsum_Wrap(x, 2)  ; result (/1  3  6  10  15/)

Example 2: When one or more _FillValue are encountered, each option yields a different answer.

  x        = (/1,2,-999,4,5/)
  x@_FillValue = -999 
  x_sum    = sum(x)            ; result 12
  x_cumsum = dim_cumsum(x, 0)  ; result (/1  3 -999 -999 -999/)
  x_cumsum = dim_cumsum(x, 1)  ; result (/1  3 -999    7   12/)
  x_cumsum = dim_cumsum(x, 2)  ; result (/1  3    3    7   12/)

  ; Use dim_cumsum_Wrap if metadata retention is desired
  ; x_cumsum = dim_cumsum_Wrap(x, 0)  ; result (/1  3 -999 -999 -999/)
  ; x_cumsum = dim_cumsum_Wrap(x, 1)  ; result (/1  3 -999    7   12/)
  ; x_cumsum = dim_cumsum_Wrap(x, 2)  ; result (/1  3    3    7   12/)

Example 3: Let p(time,lat,lon), use dimension reordering to make "time" the rightmost dimension and calculate cumulative sums at each time. The opt argument can be set by the user.

  p_cumsum = dim_cumsum(p(lat|:,lon|:,time|:), opt)  ; (nlat,mlon,ntim)

  ; Use dim_cumsum_Wrap if metadata retention is desired
  p_cumsum = dim_cumsum_Wrap(p(lat|:,lon|:,time|:), opt)

Note: Use dim_cumsum_n to avoid having to reorder your data.