pandas之DataFrame使用详解——看完不会用你打我

DataFrame是Pandas中的一个表格型数据结构，包含一组有序的列，每列值的类型都可不同（整型、浮点型、布尔型、字符串等），DataFrame既有行索引也有列索引，行标签index默认是序号 0,1,2我们也可以手动对其赋值，示例如下：

index(隐式列)	Chinese	Math	English
查良镛	99	12	13
熊耀华	98	22	23
陈文统	97	32	33

一、DataFrame实例化

DateFrame可以使用字典、列表等进行实例化，后面根据增删改查功能分类进行代码展示，示例代码均已通过测试并添加了相关注释，看官可放心使用。增删改查功能代码中使用的变量就是本步骤中所创建的df01、df02变量。

import pandas as pd, numpy as np

df01 = pd.DataFrame(data=[[11,12,13],[21,22,23],[31,32,33],[41,None,None]], 
                    columns=['Chinese','Math','English'], 
                    index=['小王','小李','小张','小空'])
df02 = pd.DataFrame({"Fruit":["apple","pear","banana","banana"], 
                     "Producer":["beijing","shanghai","taiwan","taiguo"], 
                     "Price":[1.0,2.0,3.0,4.0],
                     "Weight":[100,200,300,400]})
print(df01)
print(df02)

if __name__ == "__main__":
    filter_dataFrame()
    #update_dataFrame()
    #insert_dataFrame()
    #delete_dataFrame()

二、过滤查询

def filter_dataFrame():
    #不指定会返回所有列（写法1）
    df02_split01 = df02.loc[df02.Fruit=="banana"][df02.Producer=="taiwan"]
    #有多个条件且要指定列时，列必须放在第一个条件中或放在最后（写法1）
    df02_split02 = df02.loc[df02.Fruit=="banana",["Fruit","Producer","Weight"]][df02.Producer=="taiwan"]
    df02_split03 = df02.loc[df02.Fruit=="banana"][df02.Producer=="taiwan"][["Fruit","Producer","Weight"]]
    #可以不设置过滤条件，仅指定需要保留的列
    df02_split04 = df02[["Fruit","Producer"]]
    print('\r\n df02_split01 as below:')
    print(df02_split01)
    print('\r\n df02_split02 as below:')
    print(df02_split02)
    print('\r\n df02_split03 as below:')
    print(df02_split03)
    print('\r\n df02_split04 as below:')
    print(df02_split04)

    #不指定会返回所有列（写法2）
    df02_split05 = df02.loc[df02["Fruit"]=="banana"][df02["Producer"]=="taiwan"]
    #有多个条件且要指定列时，列必须放在第一个条件中或放在最后（写法2）
    df02_split06 = df02.loc[df02["Fruit"]=="banana",["Fruit","Producer","Weight"]][df02["Producer"]=="taiwan"]
    df02_split07 = df02.loc[df02["Fruit"]=="banana"][df02["Producer"]=="taiwan"][["Fruit","Producer","Weight"]]
    print('\r\n df02_split05 as below:')
    print(df02_split05)
    print('\r\n df02_split06 as below:')
    print(df02_split06)
    print('\r\n df02_split07 as below:')
    print(df02_split07)

三、更新数据

def update_dataFrame():
    df02.loc[df02["Fruit"]=="banana", "Price"] += 0.1
    print('\r\n df02 as below:')
    print(df02)

    df02_split01 = df02.loc[df02["Fruit"]=="banana", ["Price", "Weight"]] = [9.0, 900]
    print('\r\n df02 as below:')
    print(df02)

    col_index = df02.loc[df02["Fruit"]=="banana"][df02["Producer"]=="taiwan"].index
    df02.loc[col_index, "Price"] = 10.0
    print('\r\n df02 as below:')
    print(df02)

    col_index = df02.loc[df02["Fruit"]=="banana"][df02["Producer"]=="taiwan"].index
    df02.loc[col_index, ["Price", "Weight"]] = [11.0, 1100]
    print('\r\n df02 as below:')
    print(df02)

四、插入数据 

def insert_dataFrame():
    #若存在index=9的行就更新本行的值,不存在就在最后添加一行
    df02.loc[9] = ["new01","new01",9.0,9.0]
    print('\r\n df02 as below:')
    print(df02)
    df02.loc[len(df02)] = ["new02","new02",9.0,9.0]
    print('\r\n df02 as below:')
    print(df02)
    #默认新列的所有值都是None
    df02['newCol'] = None
    print('\r\n df02 as below:')
    print(df02)
    #如果需要为所有新列手动赋值时，数组元素的个数与DataFrame数据行的个数必须一致
    df02['newCo2'] = [1,1,1,1,1,1]
    print('\r\n df02 as below:')
    print(df02)

五、删除数据

def delete_dataFrame():
    '''
    drop()方法可用于删除DataFrame中的行或列，它有两个重要的参数：labels、axis、inplace
    -labels参数用于指定要删除的行或列的标签。可以传递单个标签，也可以传递一个标签列表
    -axis参数用于指定删除行还是列。默认值0表示删除行，1表示删除列
    -inplace参数用于指定是否在原始DataFrame上进行修改。值为True会在原始DataFrame上进行修改，值为False或不传此参数将返回一个新的DataFrame，而不修改原始DataFrame
    '''
    #删除行
    df02_split01 = df02.drop(0, axis=0, inplace=False)
    print('\r\n inplace=False, drop 0 df02 as below:')
    print(df02)
    print('\r\n inplace=False, drop 0 df02_split01 as below:')
    print(df02_split01)
    df02.drop(3, axis=0, inplace=True)
    print('\r\n inplace=True, drop 3 df02 as below:')
    print(df02)
    df02.drop([1,2], axis=0, inplace=True)
    print('\r\n inplace=True, drop 1、2 df02 as below:')
    print(df02)
    #删除列
    df02_split02 = df02.drop("Price", axis=1, inplace=False)
    print('\r\n inplace=False, drop Price df02 as below:')
    print(df02)
    print('\r\n inplace=False, drop Price df02_split01 as below:')
    print(df02_split02)
    df02.drop(["Producer","Weight"], axis=1, inplace=True)
    print('\r\n inplace=True, drop Producer、Weight df02 as below:')
    print(df02)

    '''
    del()方法可用于删除DataFrame中的列
    '''
    del df02["Price"]
    print('\r\n after del Price df02 as below:')
    print(df02)

    '''
    使用布尔索引删除行/过滤行
    '''
    #过滤出"Chinese"列值大于11的行，原始DataFrame不变
    df01_split01 = df01[df01["Chinese"]>11]
    print('\r\n filter Chinese>11, df01 as below:')
    print(df01)
    print('\r\n filter Chinese>11, df01_split01 as below:')
    print(df01_split01)

    #过滤出"Chinese"列值小于等于11的行，原始DataFrame不变
    df01_split02 = df01[~(df01["Chinese"]>11)]
    print('\r\n filter Chinese<=11, df01 as below:')
    print(df01)
    print('\r\n filter Chinese<=11, df01_split02 as below:')
    print(df01_split02)

    '''
    使用query()删除行/过滤行
    '''
    df01_split03 = df01.query("Chinese>11 & Math>22")
    df01_split04 = df01.query("Chinese>11 and Math>22")
    print('\r\n query Chinese>11 &/and Math>22, df01 as below:')
    print(df01)
    print('\r\n query Chinese>11 & Math>22, df01_split03 as below:')
    print(df01_split03)
    print('\r\n query Chinese>11 and Math>22, df01_split04 as below:')
    print(df01_split04)

    '''
    使用dropna()删除DataFrame中包含缺失值的行或列
    '''
    #删除包含缺失值的行
    df01_split05 = df01.dropna(axis=0, inplace=False)
    print('\r\n df01_split05 as below:')
    print(df01_split05)
    #删除包含缺失值的列
    df01_split06 = df01.dropna(axis=1, inplace=False)
    print('\r\n df01_split06 as below:')
    print(df01_split06)

    '''
    使用fillna()填充DataFrame中的缺失值
    '''
    df01_split07 = df01.fillna(0, inplace=False)
    print('\r\n df01_split07 as below:')
    print(df01_split07)