dataframe中 常用的 方法

新建dataframe

import pandas as pd
a = pd.DataFrame([[1,2,3],
                  [4,5,6],
                  [7,8,9]],columns = [\"feature_1\", \"feature_2\", \"label\"])

读取
import pandas as pd
df = pd.read_csv(\"datas/hour.csv\", sep=\",\")

删除dataframe列

del df[\"instant\"]

df.drop(columns=[\"instant\",\"dteday\"])

修改dataframe列名

暴力
a.columns = [\'a\',\'b\',\'c\']

较好的方法
a.rename(columns={\'A\':\'a\', \'B\':\'b\', \'C\':\'c\'}, inplace = True)

查看dataframe字段信息

a.info()

修改dataframe列类型

df[\"instant\"] = df[\"instant\"].astype(\"object\")
X[[\'Global_active_power\',\"b\"]] = X[[\'Global_active_power\',\"b\"]].astype(\'float64\')

查看dataframe统计信息

a.describe()

获取dataframe部分列

a.iloc[:,0:3]
df.iloc[:,[-1]]
a[[\"feature_1\", \"feature_2\"]]

获取dataframe列名

df.columns返回一个可迭代对象
for i in df.columns:
    print(i)

获取dataframe的Series

一行
a.iloc[0,:]

一列
a.iloc[:,1]
a[\"feature_1\"]

合并dataframe

横向
pd.concat([a,a],axis=1)

纵向
pd.concat([a,a],axis=0)

替换DF中的字符串

#df.int_rate.replace(\'%\',\'\',inplace = True, regex = True)
a.replace(\'%\',\'\',inplace = True, regex = True)

Dataframe copy

import pandas as pd
a = pd.DataFrame([[1,2,3],
                  [4,5,6],
                  [7,8,9]],columns = [\"feature_1\", \"feature_2\", \"label\"])
b = a.copy()
b.drop(columns=[\"feature_1\"],inplace=True)
a