用熊猫消除循环

Question

我有一个数据框架

ifa_num=[0.1,0.2,0.3,0.4,0.5]
ak_num=[0.6,0.7,0.8,0.9,0.11]
ch_dist=['if','ak','if','if','ak']
df=pd.DataFrame()
df['if_num']=ifa_num
df['ak_num']=ka_num
df['ch_dist']=ch_dist

点击这里看

我必须插入另一栏，即

• if_ak:如果它的ch_dist是' if‘，它将有来自ifa_num的号码，否则它将从ak_num获得号码。

生成的数据框架应该如下所示

ak柱

我编写了一个简单的代码，使用for循环，我无法弄清楚如何优化它，因为在数据增长的循环中，循环会变得效率低下。

li=[]    
for x in range(df.shape[0]):
    if df.loc[x,'ch_dist']=='if':
        li.append(df.loc[x,'if_num'])
    else:
        li.append(df.loc[x,'ak_num'])
df['if_ak']=li

Answer 1

试试下面的代码：

ifa_num=[0.1,0.2,0.3,0.4,0.5]
ak_num=[0.6,0.7,0.8,0.9,0.11]
ch_dist=['if','ak','if','if','ak']
df=pd.DataFrame()
df['if_num']=ifa_num
df['ak_num']=ak_num
df['ch_dist']=ch_dist

而不是for循环

df.loc[df['ch_dist'] == 'if', 'if_ak'] = df['if_num']
df.loc[df['ch_dist'] != 'if', 'if_ak'] = df['ak_num']

Answer 2

您可以使用布尔掩码来乘法相应的列：

import pandas as pd

ifa_num=[0.1,0.2,0.3,0.4,0.5]
ak_num=[0.6,0.7,0.8,0.9,0.11]
ch_dist=['if','ak','if','if','ak']
df=pd.DataFrame({'if_num': ifa_num,
                 'ak_num': ak_num,
                 'ch_dist': ch_dist})

m_if = df['ch_dist'] == 'if'
df['if_ak'] = m_if * df['if_num'] + (1-m_if) * df['ak_num']

df
   if_num  ak_num ch_dist  if_ak
0     0.1    0.60      if   0.10
1     0.2    0.70      ak   0.70
2     0.3    0.80      if   0.30
3     0.4    0.90      if   0.40
4     0.5    0.11      ak   0.11