numpy.cumsum#

numpy.cumsum(a,axis=None,dtype=None,out=None)[source]#

Return the cumulative sum of the elements along a given axis.

Parameters:
aarray_like

Input array.

axisint, optional

Axis along which the cumulative sum is computed. The default(None) is to compute the cumsum over the flattened array.

dtypedtype, optional

Type of the returned array and of the accumulator in which theelements are summed. Ifdtype is not specified, it defaultsto the dtype ofa, unlessa has an integer dtype with aprecision less than that of the default platform integer. Inthat case, the default platform integer is used.

outndarray, optional

Alternative output array in which to place the result. It musthave the same shape and buffer length as the expected outputbut the type will be cast if necessary. SeeOutput type determinationfor more details.

Returns:
cumsum_along_axisndarray.

A new array holding the result is returned unlessout isspecified, in which case a reference toout is returned. Theresult has the same size asa, and the same shape asa ifaxis is not None ora is a 1-d array.

See also

cumulative_sum

Array API compatible alternative forcumsum.

sum

Sum array elements.

trapezoid

Integration of array values using composite trapezoidal rule.

diff

Calculate the n-th discrete difference along given axis.

Notes

Arithmetic is modular when using integer types, and no error israised on overflow.

cumsum(a)[-1] may not be equal tosum(a) for floating-pointvalues sincesum may use a pairwise summation routine, reducingthe roundoff-error. Seesum for more information.

Examples

>>>importnumpyasnp>>>a=np.array([[1,2,3],[4,5,6]])>>>aarray([[1, 2, 3],       [4, 5, 6]])>>>np.cumsum(a)array([ 1,  3,  6, 10, 15, 21])>>>np.cumsum(a,dtype=float)# specifies type of output value(s)array([  1.,   3.,   6.,  10.,  15.,  21.])
>>>np.cumsum(a,axis=0)# sum over rows for each of the 3 columnsarray([[1, 2, 3],       [5, 7, 9]])>>>np.cumsum(a,axis=1)# sum over columns for each of the 2 rowsarray([[ 1,  3,  6],       [ 4,  9, 15]])

cumsum(b)[-1] may not be equal tosum(b)

>>>b=np.array([1,2e-9,3e-9]*1000000)>>>b.cumsum()[-1]1000000.0050045159>>>b.sum()1000000.0050000029
On this page