Pandas support

There are 2 ways pandas can handle physipy:

1. Basic support where it stores each value in a single Quantity : works out-of-the-box but with low performance
2. using physipandas : a package that takes care of the interface between phyisipy and pandas

See physipandas for better interface between pandas and physipy.

Basic support out-of-the-box

Without anything else, physipy is kinda supported in pandas, but performances will be quite degraded. A 1d quantity array will be split element-wise and stored as a squence of scalar quantities, hence all operations will be done "loop"-wise, loosing the power of numpy arrays.

import numpy as np
import pandas as pd

import physipy
from physipy import K, s, m, kg, units

# create sample data as regular Quantity
arr = np.arange(10)
heights = np.random.normal(1.8, 0.1, 10)*m
heights.favunit = units["mm"]
weights = np.random.normal(74, 3, 10)*kg

print(weights)
print(heights)

[68.02330891 75.29084364 78.31964794 77.95388057 75.15770254 72.83401712
 78.2871326  75.31371424 73.74678185 72.86442501] kg
[1905.11404675 1793.50534376 1645.47961003 1752.28899466 1705.10773302
 1816.83476718 1930.41143909 1772.03471014 1764.50051484 1921.09507113] mm

# then store Quantity in a DataFrame
df = pd.DataFrame({
    "heights":heights,
    "temp":arr*K,
    "weights":weights, 
    "arr":arr,
})
# notice the warnings below : Quantities are converted back to regular numpy arrays
# hence loosing their units.
df["heights"]

C:\Users\ym\Documents\REPOS\physipy\physipy\quantity\quantity.py:753: UserWarning: The unit of the quantity is stripped for __array__
  warnings.warn(f"The unit of the quantity is stripped for {item}")
C:\Users\ym\Documents\REPOS\physipy\physipy\quantity\quantity.py:753: UserWarning: The unit of the quantity is stripped for __array_struct__
  warnings.warn(f"The unit of the quantity is stripped for {item}")





0    1.773872
1    1.650219
2    1.876401
3    1.815951
4    1.803662
5    1.832865
6    1.776934
7    1.908975
8    1.918552
9    1.815539
Name: heights, dtype: float64

Full support using physipandas

You can also make pandas handle physipy quantities almost transparently using physipandas, which is another package that extends physipy capabilities to pandas.

Previously as part of the core project, physipandas has been moved to its own repo in order to keep phyisipy as lightweight and simple as possible.