如何比较两个数据帧列？

Question

import pandas as pd
import quandl
import numpy as np
import matplotlib.pyplot as plt
from matplotlib import style
style.use("fivethirtyeight")
df_2010=pd.read_csv("c:/users/ashub/downloads/documents/MLB 2010.csv",index_col=0)
#print(df_2010)
sliced_data=df_2010[["Home Team","Away Team","Home Score","Away Score"]]
#print(sliced_data)
for win in sliced_data:
    flag1=sliced_data["Home Team"]+str("index")
    flag2=sliced_data["Away Team"]+str("index")
    print(sliced_data["Home Score"],sliced_data["Away Score"])
    if sliced_data["Home Score"]>sliced_data["Away Score"]:
        df_2010=df_2010.join([1,0],index=[flag1,flag2])
    else:
        df_2010=df_2010.join([0,1],index=[flag1,flag2])
df_2010.to_html("c:/users/ashub/desktop/ashu.html")

ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

The error is at if condition when i am comparing the score of home team and away team.What I want to do is to add a column to the csv file which lists the win or loss of a team,win being 1 and loss being zero so that i can add the win of a particular team in a season and calculate their probability of winning and predict the probability of winning in the next season,

Answer 1

You can do just this:

df_2010['Win'] = df_2010['Home Score'] > df_2010['Away Score']

You won't need that sliced data frame.

Here's a full example:

import pandas as pd
import numpy as np

df = pd.DataFrame([np.random.randint(0, 5, 5), 
                   np.random.randint(0, 5, 5)], 
                  index=['Home Score', 'Away Score']).T

print(df)

df['Win'] = df['Home Score'] > df['Away Score']

print(df)

Which will add to

   Home Score  Away Score
0           3           3
1           4           2
2           4           1
3           4           4
4           4           2

an additional column win like this:

   Home Score  Away Score    Win
0           3           3  False
1           4           2   True
2           4           1   True
3           4           4  False
4           4           2   True

Answer 2

I think you can create boolean mask by compare columns and then assign new columns:

np.random.seed(123)
sliced_data = pd.DataFrame([np.random.randint(0, 5, 5), 
                   np.random.randint(0, 5, 5)], 
                  index=['Home Score', 'Away Score']).T

m = sliced_data['Home Score'] > sliced_data['Away Score']


sliced_data['Away Team index'] = (~m).astype(int)
sliced_data['Home Team index'] = m.astype(int)

print(sliced_data)
   Home Score  Away Score  Away Team index  Home Team index
0           2           2                1                0
1           4           3                0                1
2           2           1                0                1
3           1           1                1                0
4           3           0                0                1

It is same as:

sliced_data['Away Team index'] = np.where(m, 0,1)
sliced_data['Home Team index'] = np.where(m, 1,0)

print(sliced_data)
   Home Score  Away Score  Away Team index  Home Team index
0           2           2                1                0
1           4           3                0                1
2           2           1                0                1
3           1           1                1                0
4           3           0                0                1