在Pandas中组合列以创建新列

Question

我有下面的数据，希望将下面的列组合成一个新的列，它是二进制的，no=0，yes=1。我想合并到新列中的功能包括：

有没有告诉过你有充血性心力衰竭，有没有告诉过你有冠心病， 有没有告诉过你有心绞痛/心绞痛，有没有告诉过你心脏病发作，有没有告诉过你中风

Age in years at screening                   15881 non-null float64
Race/Hispanic origin                        15881 non-null object
Ratio of family income to poverty           15881 non-null float64
Gender                                      15881 non-null object
year                                        15881 non-null object
60 sec. pulse (30 sec. pulse * 2)           15881 non-null float64
Weight (kg)                                 15881 non-null float64
Standing Height (cm)                        15881 non-null float64
Waist Circumference (cm)                    15881 non-null float64
Arm Circumference (cm)                      15881 non-null float64
Ever told had congestive heart failure      15881 non-null object
Ever told you had coronary heart disease    15881 non-null object
Ever told you had angina/angina pectoris    15881 non-null object
Ever told you had heart attack              15881 non-null object
Ever told you had a stroke                  15881 non-null object
Do you now smoke cigarettes?                15881 non-null object
Doctor told you have diabetes               15881 non-null object
How often drink alcohol over past 12 mos    15881 non-null float64
Sodium (mmol/L)                             15881 non-null float64
Cholesterol, refrigerated serum (mg/dL)     15881 non-null float64
avg_systolic_blood_pres                     15881 non-null float64
avg_diastolic_blood_pres                    15881 non-null float64

我还担心最终可能会得到比原始数据集更多的数据（15881行，22列）

Answer 1

假设您的数据是这种格式（表是转置的，零是虚拟变量）

                                          15881  15882  15883
Q                                                            
Age_in_years_at_screening                     0      0      0
Race/Hispanic_origin                          0      0      0
Ratio_of_family_income_to_poverty             0      0      0
Gender                                        0      0      0
year                                          0      0      0
60_sec._pulse_(30_sec._pulse_*_2)             0      0      0
Weight_(kg)                                   0      0      0
Standing_Height_(cm)                          0      0      0
Waist_Circumference_(cm)                      0      0      0
Arm_Circumference_(cm)                        0      0      0
Ever_told_had_congestive_heart_failure    False  False  False
Ever_told_you_had_coronary_heart_disease   True  False  False
Ever_told_you_had_angina/angina_pectoris   True  False   True
Ever_told_you_had_heart_attack             True  False   True
Ever_told_you_had_a_stroke                 True  False   True
Do_you_now_smoke_cigarettes?                  0      0      0
Doctor_told_you_have_diabetes                 0      0      0
How_often_drink_alcohol_over_past_12_mos      0      0      0
Sodium_(mmol/L)                               0      0      0
Cholesterol_refrigerated_serum_(mg/dL)        0      0      0
avg_systolic_blood_pres                       0      0      0
avg_diastolic_blood_pres                      0      0      0

您可以指定感兴趣的问题并对其进行操作

questions = ['Ever_told_had_congestive_heart_failure',
'Ever_told_you_had_coronary_heart_disease',
'Ever_told_you_had_angina/angina_pectoris',
'Ever_told_you_had_heart_attack',
'Ever_told_you_had_a_stroke']

df["Ever_told_combined"] = df[questions].apply(lambda row: np.logical_or.reduce(row), axis=1)

这会将列“Ever\u tell\u combined”添加到数据帧中

15881     True
15882    False
15883     True
dtype: bool

Answer 2

如果要创建一个新列，如果其中任何列有“1”，则该列返回“true”，可以执行以下操作：

df = pd.DataFrame({'congestive': np.random.randint(2, size=10),
                   'coronary': np.random.randint(2, size=10)})



df['new'] = (df['congestive'] == 1) | (df['coronary'] == 1)

Out[66]: 
   congestive  coronary    new
0           1         1   True
1           1         1   True
2           1         0   True
3           1         1   True
4           0         0  False
5           0         0  False
6           0         1   True
7           1         0   True
8           0         1   True
9           1         1   True

有关将真/假更改为1/0的信息，请参见Is there a simple way to change a column of yes/no to 1/0 in a Pandas dataframe?