作者：俊欣

來源：關于數據分析與可視化

今天我們繼續來講一下Pandas和SQL之間的聯用，我們其實也可以在Pandas當中使用SQL語句來篩選數據，通過Pandasql模塊來實現該想法，首先我們來安裝一下該模塊

pip install pandasql

要是你目前正在使用jupyter notebook，也可以這么來下載

!pip install pandasql 

導入數據

我們首先導入數據

import pandas as pd from pandasql import sqldf
df = pd.read_csv("Dummy_Sales_Data_v1.csv", sep=",")
df.head()

output

我們先對導入的數據集做一個初步的探索性分析，

df.info()

output

<class 'pandas.core.frame.DataFrame'> RangeIndex: 9999 entries, 0 to 9998 Data columns (total 12 columns):
 #   Column               Non-Null Count  Dtype  
---  ------               --------------  ----- 0 OrderID 9999 non-null int64 1 Quantity 9999 non-null int64 2 UnitPrice(USD) 9999 non-null int64 3 Status 9999 non-null object 4 OrderDate 9999 non-null object 5 Product_Category 9963 non-null object 6 Sales_Manager 9999 non-null object 7 Shipping_Cost(USD) 9999 non-null int64 8 Delivery_Time(Days) 9948 non-null float64 9 Shipping_Address 9999 non-null object 10 Product_Code 9999 non-null object 11 OrderCode 9999 non-null int64  
dtypes: float64(1), int64(5), object(6)
memory usage: 937.5+ KB

再開始進一步的數據篩選之前，我們再對數據集的列名做一個轉換，代碼如下

df.rename(columns={"Shipping_Cost(USD)":"ShippingCost_USD", "UnitPrice(USD)":"UnitPrice_USD", "Delivery_Time(Days)":"Delivery_Time_Days"},
          inplace=True)
df.info()

output

<class 'pandas.core.frame.DataFrame'> RangeIndex: 9999 entries, 0 to 9998 Data columns (total 12 columns):
 #   Column              Non-Null Count  Dtype  
---  ------              --------------  ----- 0 OrderID 9999 non-null int64 1 Quantity 9999 non-null int64 2 UnitPrice_USD 9999 non-null int64 3 Status 9999 non-null object 4 OrderDate 9999 non-null object 5 Product_Category 9963 non-null object 6 Sales_Manager 9999 non-null object 7 ShippingCost_USD 9999 non-null int64 8 Delivery_Time_Days 9948 non-null float64 9 Shipping_Address 9999 non-null object 10 Product_Code 9999 non-null object 11 OrderCode 9999 non-null int64  
dtypes: float64(1), int64(5), object(6)
memory usage: 937.5+ KB

用SQL篩選出若干列來

我們先嘗試篩選出OrderID、Quantity、Sales_Manager、Status等若干列數據，用SQL語句應該是這么來寫的

SELECT OrderID, Quantity, Sales_Manager, 
Status, Shipping_Address, ShippingCost_USD 
FROM df 

與Pandas模塊聯用的時候就這么來寫

query = "SELECT OrderID, Quantity, Sales_Manager,
Status, Shipping_Address, ShippingCost_USD 
FROM df" df_orders = sqldf(query) df_orders.head() 

output

SQL中帶WHERE條件篩選

我們在SQL語句當中添加指定的條件進而來篩選數據，代碼如下

query = "SELECT * 
        FROM df_orders 
        WHERE Shipping_Address = 'Kenya'" df_kenya = sqldf(query) df_kenya.head() 

output

而要是條件不止一個，則用AND來連接各個條件，代碼如下

query = "SELECT *  FROM df_orders  WHERE Shipping_Address = 'Kenya'  AND Quantity < 40  AND Status IN ('Shipped', 'Delivered')"
df_kenya = sqldf(query)
df_kenya.head() 

output

分組

同理我們可以調用SQL當中的GROUP BY來對篩選出來的數據進行分組，代碼如下

query = "SELECT Shipping_Address,  COUNT(OrderID) AS Orders  FROM df_orders  GROUP BY Shipping_Address"

df_group = sqldf(query)
df_group.head(10) 

output

排序

而排序在SQL當中則是用ORDER BY，代碼如下

query = "SELECT Shipping_Address,  COUNT(OrderID) AS Orders  FROM df_orders  GROUP BY Shipping_Address  ORDER BY Orders"

df_group = sqldf(query)
df_group.head(10) 

output

數據合并

我們先創建一個數據集，用于后面兩個數據集之間的合并，代碼如下

query = "SELECT OrderID,
        Quantity, 
        Product_Code, 
        Product_Category, 
        UnitPrice_USD 
        FROM df" df_products = sqldf(query) df_products.head() 

output

我們這里采用的兩個數據集之間的交集，因此是INNER JOIN，代碼如下

query = "SELECT T1.OrderID, 
        T1.Shipping_Address, 
        T2.Product_Category 
        FROM df_orders T1
        INNER JOIN df_products T2
        ON T1.OrderID = T2.OrderID" df_combined = sqldf(query) df_combined.head() 

output

與LIMIT之間的聯用

在SQL當中的LIMIT是用于限制查詢結果返回的數量的，我們想看查詢結果的前10個，代碼如下

query = "SELECT OrderID, Quantity, Sales_Manager,  Status, Shipping_Address, 
ShippingCost_USD FROM df LIMIT 10"

df_orders_limit = sqldf(query)
df_orders_limit 

output