给定两个具有相同列的DataFrame.
DFA
Index Price 0 10.21 1 12.21
DFB
Index Price 0 10.21 1 10.24 2 11.32 3 12.21
I want to add a column to DFA with the times that each value appears in DFB, but with a tolerance of, let's say, 1%.
结果
Index Price Occurrences 0 10.21 2 1 12.21 1
是否仍有可能避免迭代？也许用merge_asof和grouping？
附注:这是对我的其他question条的修正，但发布了不同的威胁，因为它涉及的是一个略有不同的问题.
As you want to test all combinations, use numpy broadcasting:
a = dfA['Price'].to_numpy()[:,None]
b = dfB['Price'].to_numpy()
dfA['Occurrences'] = (abs(a-b) < a*0.01).sum(axis=1)
输出:
Index Price Occurrences
0 0 10.21 2
1 1 12.21 1
中间体:
# abs(a-b)
array([[0. , 0.03, 1.11, 2. ],
[2. , 1.97, 0.89, 0. ]])
# (abs(a-b) < a*0.01)
array([[ True, True, False, False],
[False, False, False, True]])