pandas常用操作

1.转换数据类型
df.apply(pd.to_numeric,errors='ignore')
2.宽数据转换为长数据
pd.melt(df,id_vars=['col3'])
3.重命名列名称
df.rename(columns={'col1':'列1'},inplace=True)
4.crosstab 混淆矩阵
pd.crosstab(df['truth'],df['predict'])
5.join操作
merge_df=pd.merge(df1,df2,on='col1',how='left')
6.重置索引
df.reset_index()
7.替换填充
df.replace(0,1,inplace=True)
df.fillna(1)
8.groupby去掉索引
df.groupby('col1',as_index=False).size()
9.one-hot编码
time_dummy = pd.get_dummies(raw_df['timety_2'])
static_df = pd.concat([time_dummy,raw_df],axis=1)
10.读取excel
pd.read_excel(io=r'test.xlsx',sheet_name='factor',engine='openpyxl')