python:pandas学习笔记

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")

#用前一个数据代替NaN：method='pad'
time_info.fillna(method='pad')
#与pad相反，bfill表示用后一个数据代替NaN
time_info.fillna(method='bfill')
#用limit限制每列可以替代NaN的数目
time_info.fillna(method='bfill',limit=1)
#使用平均数代替NaN
time_info.fillna(time_info.mean())
#指定列 数据代替NaN
time_info.fillna(time_info.mean()['SUBTIME':'OPRTIME'])

test_info = sub_info[["CONTRACTID","STATUS"]]

#A value is trying to be set on a copy of a slice from a DataFrame
test_num = test_info["STATUS"] / 10
test_info["test"] = test_num
test_info

test_num = sub_info["STATUS"] / 10
sub_info["test"] = test_num
sub_info

CONTRACTID_MAX = sub_info["CONTRACTID"].max()
max_percent = sub_info["CONTRACTID"] / CONTRACTID_MAX
sub_info["max_percent"] = max_percent

#对列排序，参数True 表示不产生拷贝，直接在原矩阵操作
sub_info.sort_values("CONTRACTID",inplace=True)
sub_info
#ascending 降序
sub_info.sort_values("CONTRACTID",inplace=True,ascending=False)
sub_info