Pandas中常用的方法

為資料分配索引，例如：

data=np.random

.randn(5)
pd.series(data, index=['a', 'b', 'c', 'd', 'e'])
>>>
a -0.287461
b 0.736157
c 1.759875
d -0.238167
e 0.621458
dtype: float64

pd.series(np.random

.randn(5))
>>>
0 -0.334205
1 -1.033102
2 -0.349577
3 -1.459086
40.148646
dtype: float64

df1 = pd.dataframe(,

index=[0, 1, 2, 3])
df2 = pd.dataframe(,
index=[0, 1, 2, 3])
s=pd.concat([df1,df2], axis=1) # 1是在x軸方向合併，0是在y軸方向合併

df.drop([『column』],axis=1,inplace=true) <—–> df = df.drop([『column』],axis=1)

df[『column』]=df[『column』].astype(『int』)

df[『column』]=df.column.astype(int)

df[『column』].fillna(value=0, inplace=true)

df[『column』][df[『column』]==value1]=value2

類似於將sql語句：select df1.id from df1, df2 where df1.id=df2.id轉換成pandas語句

df1[df1['id']==df2['id']]這個辦法表df1, df2必須有相同的index

否則會出現

valueerror: can only compare identically-labeled series objects

錯誤

col_dates = df.dtypes[df.dtypes == 'datetime64[ns]'].index

for d in col_dates:
df[d] = df[d].dt.to_period('m')

df['emp_length'] = df['emp_length'].fillna(df.emp_length.median())

Pandas常用的方法

讀取 寫入read csv to csv read excel to excel read hdf to hdf read sql to sql read json to json read msgpack experimental to msgpack experimental read html...

pandas常用方法

import pandas as pd import numpy as np import matplotlib.pyplot as plt import datetime import redf pd.read csv path file.csv 引數 header none 用預設列名，0，1，...

pandas 常用方法

import pandas as pd pd.read csv filename,encoding utf 8 讀取csv pd.to csv filename 儲存檔案，filename為檔案路徑，可以是相對路徑or絕對路徑 pd.to csv filename,index 0 儲存到檔案時，不要...