Python Pandas：如何从另一个csv文件更新一个csv文件

Question

We have two CSV files: a.csv and b.csv .

a.csv has tree columns: label , item1 , item2 . b.csv has two columns: item1 , item2 . If item1 and item2 in a.csv also occurr in b.csv , that's a.csv and b.csv have same item1 and item2 , the value of label in a.csv should be 1 instead. How to use pandas to deal?

---

For example:

a.csv:

label    item1     item2
 0         123       35
 0         342       721
 0         876       243

b.csv:

item1     item2
 12        35
 32        721
 876       243

result.csv:

label    item1     item2
 0         123       35
 0         342       721
 1         876       243

I tried this, but it doesn't work:

import pandas as pd

df1 = pd.read_csv("~/train_dataset.csv", names=['label', 'user_id', 'item_id', 'behavior_type', 'user_geohash', 'item_category', 'time','sales'], parse_dates=True)
df2 = pd.read_csv(~/train_user.csv", names=['user_id', 'item_id', 'behavior_type', 'user_geohash', 'item_category', 'time', 'sales'], parse_dates=True)
df1.loc[(df1['user_id'] == df2['user_id'])& (df1['item_id'] == df2['item_id']), 'label'] = 1

Answer 1

You could use loc and a boolean condition to mask your df (here representing a.csv) and set the label to 1 if that condition is met:

In [18]:

df.loc[(df['item1'] == df1['item1'])& (df['item2'] == df1['item2']), 'label'] = 1
df
Out[18]:
   label  item1  item2
0      0    123     35
1      0    342    721
2      1    876    243

If you want to set all row values you could use np.where :

In [19]:

np.where((df['item1'] == df1['item1'])& (df['item2'] == df1['item2']), 1, 0)
Out[19]:
array([0, 0, 1])
In [20]:

df['label'] = np.where((df['item1'] == df1['item1'])& (df['item2'] == df1['item2']), 1, 0)
df
Out[20]:
   label  item1  item2
0      0    123     35
1      0    342    721
2      1    876    243