Data structure/algorithm to dynamically change "tree"-type dependency-CodePudding

I am looking for the data structure/algorithm, which could help me to solve the following task:

imagine we have two business lines B1 and B2
each business line has 3 products P1, P2, P3
each product has same nomenclature of parameters, which my programme eventually will compute: H, W, L

To answer the comment: the values for H, W, L are indeed numbers.

To clarify further: all B, P and H,W,L are strings, but the values of H,W,L (I didn't show them on the tree) are numbers and all different.

In the first run, I was thinking about building a tree like this:

Then, I would like to perform an analytics, on let's say P1, so I need to be able to easily retrieve data from P1 node, across the tree like this:

Subsequently, analytics on H would require retrieving data like this:

(For inspiration, this is quite easily achieved in Excel pivots, by just rearranging columns)

My language is Python and so far I came up only with nested dictionaries (though it's very tedious to search for nested keys), or dictionary with B1, B2 as keys and let's say (P1, H, value_of_H) as tuples, where I can see for each key if either P or H/W/L is in tuple.

CodePudding user response：

You could create a Pandas DataFrame with multiple index levels, and query it using xs:

import pandas as pd
import numpy as np

firms = ["B1", "B2"]
products = ["P1", "P2", "P3"]
params = ["H", "W", "L"]

df = pd.DataFrame(
    index=pd.MultiIndex.from_product([firms, products, params]),
    data=np.random.randint(1, 10, 2 * 3 * 3),
    columns=["Value"]
    )

# analytics on P1:
print(df.xs("P1", level=1))
#       Value
# B1 H      5
#    W      6
#    L      9
# B2 H      1
#    W      5
#    L      7

# analytics on H:
print(df.xs("H", level=2))
#        Value
# B1 P1      7
#    P2      2
#    P3      6
# B2 P1      5
#    P2      6
#    P3      6