def select_data2collision():
data1 = data.iloc[:,45:] #取45列以后的数据
data2 = pd.concat([data["CASEID"],data1], axis=1) #把caseid的列合并进去
print(data2.head())
data3 = data2.drop_duplicates(['CASEID']) #按照cased去重
print("################")
print(data3.head())
data4 = data3.reset_index(drop=True) #重建索引
print(data4.head())