Python系列中的数据比较

这是两个数据系列，我希望在第三个数据系列的基础上进行比较

数据单元1：

SKU     Weight1     
1234    20
1235    30
111     40
101     23

数据单元2：

SKU     Weight2
1234    22
1235    35
111     47
101     87

标志数据：

SKU
1234     True
1235     False
111      True
101      False
Name: Date, dtype: bool

基本上基于flag_数据系列中的值，我需要将Weight1的值除以Weight2，反之亦然

对于intance：

j = flag_Data(dados_SKU1, dados_SKU1) #this generates the third series
if(j[1234]==True):
  generated_serie = data_SKU1['Weight1'][1234] / dados_SKU2['Weight2'][1234]
else:
  generated_serie = data_SKU2['Weight2'][1234] / data_SKU1['Weight1'][1234]

---

但应针对该系列中的所有SKU执行此操作，而不仅仅是SKU 1234。你们能帮我弄清楚怎么做吗？

你们可以用np.where来做这个：

result = np.where(flag_Data,
                  data_SKU1['Weight1']/data_SKU2['Weight2'],
                  data_SKU2['Weight2']/data_SKU1['Weight1'])

设置

合并

df = df1.merge(df2)

    SKU  Weight1  Weight2   FLAG
0  1234       20       22   True
1  1235       30       35  False
2   111       40       47   True
3   101       23       87  False

df['division'] = np.where(df['FLAG'], df['Weight1']/df['Weight2'], df['Weight2']/df['Weight1'])

df.loc[df['FLAG'], 'division'] = df.Weight1 / df.Weight2
df['division'] = df.division.fillna(df.Weight2/df.Weight1)

df['division'] = (df.Weight1 / df.Weight2.mask(~df.FLAG)).fillna(df.Weight2/df.Weight1)

选项1

np.其中

df = df1.merge(df2)

    SKU  Weight1  Weight2   FLAG
0  1234       20       22   True
1  1235       30       35  False
2   111       40       47   True
3   101       23       87  False

df['division'] = np.where(df['FLAG'], df['Weight1']/df['Weight2'], df['Weight2']/df['Weight1'])

df.loc[df['FLAG'], 'division'] = df.Weight1 / df.Weight2
df['division'] = df.division.fillna(df.Weight2/df.Weight1)

df['division'] = (df.Weight1 / df.Weight2.mask(~df.FLAG)).fillna(df.Weight2/df.Weight1)

选项2

loc

与

fillna

df = df1.merge(df2)

    SKU  Weight1  Weight2   FLAG
0  1234       20       22   True
1  1235       30       35  False
2   111       40       47   True
3   101       23       87  False

df['division'] = np.where(df['FLAG'], df['Weight1']/df['Weight2'], df['Weight2']/df['Weight1'])

df.loc[df['FLAG'], 'division'] = df.Weight1 / df.Weight2
df['division'] = df.division.fillna(df.Weight2/df.Weight1)

df['division'] = (df.Weight1 / df.Weight2.mask(~df.FLAG)).fillna(df.Weight2/df.Weight1)

选项3

掩码

使用

填充

df = df1.merge(df2)

    SKU  Weight1  Weight2   FLAG
0  1234       20       22   True
1  1235       30       35  False
2   111       40       47   True
3   101       23       87  False

df['division'] = np.where(df['FLAG'], df['Weight1']/df['Weight2'], df['Weight2']/df['Weight1'])

df.loc[df['FLAG'], 'division'] = df.Weight1 / df.Weight2
df['division'] = df.division.fillna(df.Weight2/df.Weight1)

df['division'] = (df.Weight1 / df.Weight2.mask(~df.FLAG)).fillna(df.Weight2/df.Weight1)

所有结果如下：

    SKU  Weight1  Weight2   FLAG  division
0  1234       20       22   True  0.909091
1  1235       30       35  False  1.166667
2   111       40       47   True  0.851064
3   101       23       87  False  3.782609