how do I make a numpy array from a list of pint quantities

Question

import numpy as np
from pint import UnitRegistry 
unit = UnitRegistry()
Q_ = unit.Quantity
a = 1.0*unit.meter
b = 2.0*unit.meter
# some calculations that change a and b 
x=np.array([a.magnitude,b.magnitude])*Q_(1.0,a.units)

will make a numpy array from variables a and b which are Pint quantities. It is somewhat crude as there is no guarantee that a and b have the same units. Is there a cleaner way to do it? Do I need to write a function?

score 2 · Accepted Answer · answered Jan 08 '23 at 06:36

2

This can now be done with the from_list method of pint.Quantity:

import pint
unit = pint.UnitRegistry()
a = 1.0*unit.meter
b = 2.0*unit.meter
print(pint.Quantity.from_list([a, b]))
>> [1.0 2.0] meter

answered Jan 08 '23 at 06:36

Simon Rochester

136
3

score 1 · Answer 2 · answered Jul 08 '15 at 02:15

You should convert to a base unit with to_base_units() before creating a numpy array from it, no need to write a function for this, a simple list comprehension will do it nicely.

If you're doing numerically heavy calculations on x you probably will want to keep it in a reference unit and use it as a raw array (without attached unit), and limit the unit conversions to the input/output phase of the program.

import numpy as np
from pint import UnitRegistry

unit = UnitRegistry()
Q_ = unit.Quantity

a = 1.0 * unit.meter
b = 2.0 * unit.meter
c = 39.37 * unit.inch

# A list with values in different units (meters and inches)
measures = [a, b, c]

# We use a comprehension list to get the magnitudes in a base unit
x = np.array([measure.to_base_units().magnitude for measure in measures])

print x * Q_(1.0, unit.meter)

>> [ 1.        2.        0.999998] meter

That's nice, if slightly longer. I like to keep the units around, because an error in units when adding two variables, indicates an error in the calculation. — pheon, Jul 08 '15 at 12:51

score 1 · Answer 3 · answered Mar 22 '19 at 13:40

1

Another option, which allows for arrays whose elements have different units, probably at the cost of some efficiency and fun is to give the array dtype='object'.

import numpy as np
from pint import UnitRegistry 
unit = UnitRegistry()
a = 1.0*unit.meter
b = 2.0*unit.meter
# some calculations that change a and b 
x=np.array([a, b], dtype='object')

answered Mar 22 '19 at 13:40

Oliver Evans

1,405
14
17

Interesting suggestion, but is there a way to first declare the array with float values then apply the element wise quantity object conversion? – user32882 Jul 23 '19 at 08:48
@user32882 one option would be `v = [3.0, 4.5, 5.7]; x = np.array([e*unit.meter for e in v], dtype=object)` – Oliver Evans Jul 23 '19 at 16:06
1

the new project `pint-pandas` is trying to implement this for `pandas` – user32882 Oct 13 '19 at 09:45

how do I make a numpy array from a list of pint quantities

3 Answers3

Linked