Lower-bound binning
Convert continuous measurements into lower integer-valued bins.
NumPy · Function
np.floor
Returns the mathematical floor of input values element-wise: the greatest integer less than or equal to each value. It returns an ndarray for array input and a scalar for scalar input.
How it works
np.floor is a universal function that applies the mathematical floor independently to each input element. For every value, it selects the greatest integer that is less than or equal to that value; array input produces an array, while scalar input produces a scalar.
Syntax
numpy.floor(x, /, out=None, *, where=True, casting='same_kind', order='K', dtype=None, subok=True [, signature])Parameters
| Name | Type | Description |
|---|---|---|
x | array_like | Input data. |
out | ndarray, None, or tuple of ndarray and None, optional | Destination in which the result is stored; may be an ndarray, None, or a tuple of ndarray and None. |
where | array_like of bool, optional | Condition broadcast over the input; where false, the output is not computed. |
casting | {‘no’, ‘equiv’, ‘safe’, ‘same_kind’, ‘unsafe’}, optional | Controls the data-type casting permitted for the operation. |
order | {‘K’, ‘A’, ‘C’, ‘F’}, optional | Controls the memory layout of the result. |
dtype | dtype, optional | Data type of the returned result. |
subok | bool, optional | If true, subclasses of ndarray are passed through. |
signature | str or tuple of ints, optional | Generalized ufunc signature, when applicable. |
Returns
ndarray or scalar: the floor of each element in x; a scalar if x is a scalar.
Key points
- Computes floor independently for every input element.
- For a value x, the result is the greatest integer that does not exceed x.
- The operation has the standard mathematical notation ⌊x⌋.
- The optional
outargument can store results in a supplied destination. - The returned array contains one floored result for each input element.
Common mistakes
- Flooring is not truncation toward zero: negative fractional values move to the next smaller integer.
Example: Floor an array element-wise
import numpy as np
values = np.array([-1.7, -0.2, 0.2, 1.7])
print(np.floor(values))
# [-2. -1. 0. 1.]Applies floor element-wise, including rounding negative nonintegers toward smaller values.
When to use it
Vectorized preprocessing
Apply a mathematical lower bound to every value in an ndarray without writing an explicit loop.
Masked or in-place-style output
Reuse an existing destination array or restrict updates with ufunc output controls.
Interview questions
QWhat does np.floor compute?
np.floor returns the floor of its input element-wise. For each scalar, that is the largest integer less than or equal to the value.
Follow-up: How does it behave for a scalar input? It returns a scalar rather than a one-element array.
QWhat are the input and output forms?
The function accepts array-like input and returns the floor of every element. An array input yields an ndarray, whereas a scalar input yields a scalar.
Follow-up: What happens with negative values? A value such as -0.2 becomes -1, because -1 is the greatest integer no larger than -0.2.
QHow can computation be directed or selectively applied?
As a ufunc, np.floor supports an optional out destination and a where mask in addition to the input. These enable selective computation and storage into a supplied array.
Follow-up: Why use where with out? Unselected positions can retain their preexisting destination values instead of being recalculated.
Related
numpy.ceilnumpy.truncnumpy.rint