覆蓋兩個熊貓Dataframes或numpy數組並創建一個鍵值字典

Question

我有兩個CSV文件的形狀可以說3x3。

文件1看起來像這樣：

-1, 2,-1 
-1,-1, 3 
-1, 8, 9 

和文件2是這樣的：

-1, 56,-1 
-1,-1, 73 
-1, 24, 100 

我的目標是覆蓋兩個陣列或DataFrame S和創建以下形式的字典：

dict = {2:56, 3:73, 8:24, 9:100} 

我可以創建一些基於循環的東西，但我想知道是否有辦法掩蓋或覆蓋數組並創建相應的字典？

Answer 1

您可以展平dataframes，放下空值，和字典/ ZIP他們：

import pandas as pd 

# Load the data, converting -1 to NaN 
dfa = pd.read_csv('file1.csv', header=None, na_values=[-1]) 
dfb = pd.read_csv('file2.csv', header=None, na_values=[-1]) 

# Flatten the dataframes 
a = dfa.values.flatten() 
b = dfb.values.flatten() 

# Remove null values and cast back to ints (if that matters) 
# Note that both are filtered according to the key data 
a = a[~np.isnan(a)].astype(int) 
b = b[~np.isnan(a)].astype(int) 

# Zip to pair the lists, then convert to a dict 
d = dict(zip(a, b))