Module reduce 

Reductions along named axes, cumulative operations and differences.

Nulls are skipped, like the NumPy nan* functions. A group without valid values gives 0 for sums, 1 for products, false for any, true for all, 0 for counts and null for all other reductions. Floats order NaN above all numbers, like Arrow, so max can return NaN. Use nan_to_null to treat NaN as missing.

Enums

Reduction

A reduction for reduce and groupby_reduce.

Functions

all

true if all values are true (or non-zero), like numpy.all.

any

true if any value is true (or non-zero), like numpy.any.

argmax

Index of the first maximum, like numpy.nanargmax.

argmin

Index of the first minimum, like numpy.nanargmin.

count

Number of non-null values, like xarray.DataArray.count.

cummax

Running maximum along dim, like numpy.maximum.accumulate. Null positions stay null.

cummin

Running minimum along dim, like numpy.minimum.accumulate. Null positions stay null.

cumprod

Cumulative product along dim, like numpy.nancumprod. Null positions stay null.

cumsum

Cumulative sum along dim, like numpy.nancumsum. Null positions stay null.

diff

The n-th discrete difference along dim, like numpy.diff.

max

Maximum, like numpy.nanmax.

mean

Mean, like numpy.nanmean. Integers give Float64.

median

Median, like numpy.nanmedian.

min

Minimum, like numpy.nanmin.

prod

Product, like numpy.nanprod. Integers multiply as Int64 or UInt64, with wrapping.

ptp

Range of values (max - min), like numpy.ptp.

quantile

Quantile with linear interpolation, like numpy.nanquantile.

reduce

Reduce a over the named axes. None reduces all axes to a scalar.

std

Standard deviation, like numpy.nanstd.

sum

Sum, like numpy.nansum. Integers sum as Int64 or UInt64, with wrapping.

var

Variance, like numpy.nanvar.