在 pandas 中使用 sqlite3 在列之间的范围内搜索

Question

我在一个问题中找到了解决问题的办法我试图根据我的情况修改它，但它没有用。在下面给出的代码中，我需要 df 来显示每个销售产品和类别的开始和结束。但它忽略了介于开始和结束期间之间的日期。正如在 01/06/2020 期间 26/03/2020 - 31/07/2020 期间的 Apple sel 屏幕上可以看到的那样，但它显示其他。我应该如何澄清 SQL 查询？

import pandas as pd
import sqlite3

dates_of_discount=pd.DataFrame({"Date_begining":['01/01/2021','01/02/2020','26/03/2020'],
                   "Date_ending":['31/12/2021', '25/02/2020', '31/07/2020'],
                   "Category":['Discount', 'Not Discount', "Discount"],
                   "d_Product":['Apple', 'Peach', "Apple"]})
purchase_dates=pd.DataFrame({"date":(["20/01/2020", "18/02/2020", "01/06/2020"]),
                          "Qty":[100, 200, 300],
                          "Price":[3.5,4, 20],
                          "p_Product":['Apple', 'Peach', "Apple"]})


conn = sqlite3.connect(':memory:')

dates_of_discount.to_sql('disc', conn, index=False)
purchase_dates.to_sql('purch', conn, index=False)

qry = '''
    select  
        purch.date Sold,
        purch.p_Product Prod,
        purch.Qty,
        purch.Price,
        Date_begining Period_Start,
        Date_ending Period_End,
        Category Output
    from
        purch join disc on
        date between Date_begining and Date_ending and
        d_Product = p_Product
    '''
df = pd.read_sql_query(qry, conn)
df

Answer 1

如评论所述，通过使用 pd.to_datetime 将日期字符串正确转换为 pandas 中的实际 datetime，SQL 连接操作应该 return 预期结果:

输入数据（带日期转换）

dates_of_discount = pd.DataFrame({
    "Date_begining": pd.to_datetime(
        ['01/01/2021','01/02/2020','26/03/2020'],
        format="%d/%m/%Y"
    ),
    "Date_ending": pd.to_datetime(
        ['31/12/2021', '25/02/2020', '31/07/2020'],
        format="%d/%m/%Y"
    ),
    "Category": ['Discount', 'Not Discount', "Discount"],
    "d_Product": ['Apple', 'Peach', "Apple"]
})

purchase_dates=pd.DataFrame({
    "date": pd.to_datetime(
        ["20/01/2020", "18/02/2020", "01/06/2020"],
        format="%d/%m/%Y"
    ),
    "Qty":[100, 200, 300],
    "Price":[3.5,4, 20],
    "p_Product":['Apple', 'Peach', "Apple"]
})

SQL站点查询

conn = sqlite3.connect(':memory:')

dates_of_discount.to_sql('disc', conn, index=False)
purchase_dates.to_sql('purch', conn, index=False)

qry = '''
    select  
        purch.date as Sold,
        purch.p_Product as Prod,
        purch.Qty,
        purch.Price,
        disc.Date_begining as Period_Start,
        disc.Date_ending as Period_End,
        disc.Category as Output
    from purch 
    join disc 
      on purch.date between disc.Date_begining and disc.Date_ending 
      and purch.p_Product = disc.d_Product
    '''
merge_df = pd.read_sql_query(qry, conn)
merge_df
#                   Sold   Prod  Qty  Price         Period_Start           Period_End        Output
# 0  2020-02-18 00:00:00  Peach  200    4.0  2020-02-01 00:00:00  2020-02-25 00:00:00  Not Discount
# 1  2020-06-01 00:00:00  Apple  300   20.0  2020-03-26 00:00:00  2020-07-31 00:00:00      Discount

顺便说一下，pandas 也可以运行类似的操作 merge 按产品和 query 或按日期过滤（reindex 和 set_axis 子集和重命名列）：

merge_df = (
    purchase_dates.merge(
        dates_of_discount, left_on="p_Product", right_on="d_Product"   
    ).query(
        "date >= Date_begining & date <= Date_ending"
    ).reset_index(drop=True)
    .reindex(
        ["date", "p_Product", "Qty", "Price", "Date_begining", "Date_ending", "Category"],
        axis = "columns"
    ).set_axis(
        ["Sold", "Prod", "Qty", "Price", "Period_Start", "Period_End", "Output"],
        axis = "columns",
        inplace = False
    )
)
merge_df_pd
#         Sold   Prod  Qty  Price Period_Start Period_End        Output
# 0 2020-06-01  Apple  300   20.0   2020-03-26 2020-07-31      Discount
# 1 2020-02-18  Peach  200    4.0   2020-02-01 2020-02-25  Not Discount

最后，根据您的评论，如果对 SQL 或 pandas 都使用产品尺寸等数字而不是日期：

，则相同的逻辑应该有效

select  
    purch.date as Sold,
    purch.p_Product as Prod,
    purch.Qty,
    purch.Price,
    disc.min_product_size,
    disc.max_product_size,
    disc.Category as Output
from purch 
join disc 
  on purch.product_size between disc.min_product_size and disc.max_product_size
  and purch.p_Product = disc.d_Product

merge_df = (
    purchase_dates.merge(
        dates_of_discount, left_on="p_Product", right_on="d_Product"   
    ).query(
        "product_size >= min_product_size & product_size <= max_product_size"
    ).reset_index(drop=True)
    .reindex(
        ["date", "p_Product", "Qty", "Price", "min_product_size", "max_product_size", "Category"],
        axis = "columns"
    ).set_axis(
        ["Sold", "Prod", "Qty", "Price", "min_product_size", "max_product_size", "Output"],
        axis = "columns",
        inplace = False
    )
)

在 pandas 中使用 sqlite3 在列之间的范围内搜索

searching in range between columns using sqlite3 in pandas

python

sql

sqlite

pandas