Python原始代碼如下:
import pandas as pd
import numpy as np
patient=pd.read_csv(r'C:\Users\Administrator\full_cohort_data2.csv')
last = np.zeros(patient.shape[0])
alist=['Fri','Sa','Th']
colum=['day_icu_intime','service_num','service_unit']
def charlson(day):
for i in range(1776):
for j in range(len(alist)):
if patient[day][i].startswith(alist[j]):
last[i]+=2
for k in range(len(colum)):
charlson(colum[k])
一直出現'numpy.int64' object has no attribute 'startswith',主要是因爲'service_num'這列爲數字,'service_unit'這一列有空值,無法使用pandas的內置的str函數
解決的辦法就是在函數中直接將所有要遍歷的列類型轉換成str,函數修改成這個樣子就可以了:
def charlson(day):
patient[day]=patient[day].astype(str)#將該列類型轉換爲str,列中的空值會被填補爲nan
for i in range(1776):
for j in range(len(alist)):
if patient[day][i].startswith(alist[j]):
last[i]+=2