Allows arbitrary slicing (subsampling) of the data.
Parameters:
key ([int | float | tuple | slice]) –
When key is a single number, it specifies a number/fraction of events
to use. Use the parameter ‘order’ to specify how to subsample
the requested number/fraction of events.
int : specifies a number of events to use
float : specifies a fraction of events to use (a number between 0 and 1)
tuple : consists of two floats, each between 0 and 1. For example, key = (0.66666, 1.0) returns the last one third of events.
slice : applies a slice. For example, key = slice(10, 1000, 20) returns events with indexes [10, 30, 50, ...]
order ([‘random’ | ‘start’ | ‘end’]) –
Specifies which events to choose. This is only relevant
when key is either an int or a float.
‘random’ : chooses the events randomly (without replacement)
‘start’ : subsamples starting from the start
‘end’ : subsamples starting from the end
auto_resize ([False | True]) – If True, attempts to automatically control indexing errors.
For example, if there are only 1000 events in the fcs sample,
but the key is set to subsample 2000 events, then an error will be raised.
However, with auto_resize set to True, the key will be adjusted
to 1000 events.
