1. 程式人生 > >python:pandas學習筆記

python:pandas學習筆記

python pandas 人工智能

import pandas sub_info = pandas.read_csv("contract.csv") #sub_info #print (sub_info) type(sub_info) #print (sub_info.dtypes) first_rows = sub_info.head(1) #print (first_rows) #print (sub_info.columns) #print (sub_info.shape) #print (sub_info.loc[1]) sub_info.loc[0:3] two_five_nine = [2,5,9] sub_info.loc[two_five_nine] id1 = sub_info["CONTRACTID"] id1 str1 = ["CONTRACTID","STATUS"] id2 = sub_info[str1] id2 sub_info.columns columns_list = sub_info.columns.tolist() time_list = [] for i in columns_list: if i.endswith("TIME"): time_list.append(i) time_info = sub_info[time_list] is_value_empty = time_info.isnull() is_value_empty time_info.fillna("0") #用前一個數據代替NaN:method=‘pad‘ time_info.fillna(method=‘pad‘) #與pad相反,bfill表示用後一個數據代替NaN time_info.fillna(method=‘bfill‘) #用limit限制每列可以替代NaN的數目 time_info.fillna(method=‘bfill‘,limit=1) #使用平均數代替NaN time_info.fillna(time_info.mean()) #指定列 數據代替NaN time_info.fillna(time_info.mean()[‘SUBTIME‘:‘OPRTIME‘])

python:pandas學習筆記