python pandas：对于列表中的每个元素，根据条件从数据帧返回一行

Question

I have two dataframes (df_key,df2_values) and would like to get rows from df_source for each row in df_key.

I can do a lookup for one key at a time but that is inefficient considering that the I have to loop through rows in df_key and then call a function to appropriate rows back from df_source.

df_source[((df_source["cond1"] == key) | (df_source["cond2"] == key)) & (df_source["cond3"] == "values") ]

Any suggestions???

Answer 1

I think you need isin by column in df_key :

df = df_source[((df_source["cond1"].isin(df_key['key'])) |
                (df_source["cond2"].isin(df_key['key']))) & (df_source["cond3"] == "values")]

Sample:

df_source = pd.DataFrame({'A':list('abcdef'),
                   'cond1':[4,5,4,5,5,4],
                   'cond2':[7,8,9,4,2,3],
                   'D':[1,3,5,7,1,0],
                   'E':[5,3,6,9,2,4],
                   'cond3':list('aaabbb')})


df_key = pd.DataFrame({
    'A': ['a','a','a','a','b','b','b','c','d'],
    'key': [5,3,6,9,2,4,1,2,3]
})

df = df_source[((df_source["cond1"].isin(df_key['key'])) | 
                (df_source["cond2"].isin(df_key['key']))) & (df_source["cond3"] == "a") ]
print (df)
   A  D  E  cond1  cond2 cond3
0  a  1  5      4      7     a
1  b  3  3      5      8     a
2  c  5  6      4      9     a

EDIT:

You need filter all unique values for keys use numpy.intersect1d , but is necessary also filter by condition:

vals = df_source.loc[df_source["cond3"] == "a", ['cond1','cond2']].values.ravel()
print (vals)
[4 7 5 8 4 9]

L = np.intersect1d(vals, df_key['key'].values).tolist()
print (L)
[4, 5, 9]