如何在BIGQUERY中查询当前和上一季度的数据

编程问答 2022-08-23

问题描述

我正在尝试获取BIGQUERY的上一季度和当前季度的记录。但是无法弄清楚查询的构造方式。

背景：在第二个表中，它基于MILESTONE字符串（本质上是一个字符串，表示为季度“ 2020_Q 1 _M0”（目标，而不是基于MILESTONE下的书面查询。我可以创建一个动态过滤器，如果当前季度从第二季度更改为第三季度，它将使用当前季度（第三季度）更新数字，并保留上一个季度Q2）

注释：SA_Date是我的日期格式“ 2020-05-24”

当前脚本：

-- 1st table pulling the current cost numbers based on current date from XYZ1 table

SELECT CASE WHEN PMD_PRODUCT_NAME LIKE '%MAN%' THEN 'MAN' 
    WHEN PC.COUNTRY ='BR' THEN 'Brazil'
    ELSE GEO.COUNTRY_DESCRIPTION
    END AS Country,PC.PRODUCT_STATUS AS Product_Status,PC.SA_DATE AS SA_Date,MILESTONE 
FROM
-- PC table
  (SELECT *,CASE WHEN SA_DATE < CURRENT_DATE() THEN 'Legacy' ELSE 'NPI' END PRODUCT_STATUS FROM [xxxx] 
  WHERE PMD_PRODUCT_STATUS ='Active') PC

-- Country since PC table has country codes and not description
  JOIN [xxxx] GEO
  ON PC.COUNTRY = GEO.COUNTRY_CODE
-- PCS table joined 
  LEFT JOIN 
    (SELECT APC_CODE,COUNTRY,KEY_CUSTOMER,PRODUCT_VARIANT,TMC AS PCS_TMC,BMC AS PCS_BMC,TMC_TARGET
    FROM [xxxx] 

-- 2nd table pulling the historical cost numbers (from XYZ2 table) based on MILESTONE string (**GOAL** instead basing on my query on MILESTONE string can I want it dynamic where if the current quarter changes from Q2 to Q3 and also retain the prevIoUs quarter which will be Q2)
 
    WHERE PMD_PRODUCT_STATUS = 'Active'
    **AND MILESTONE LIKE ('prevIoUs_qtr_modified') -- filtered to specific qtr**
    AND NOT (PMD_PROGRAM_NAME LIKE '%CANCELED%')
    AND LOCKED IS false
    ) PCS
    ON PC.APC_CODE = PCS.APC_CODE
    AND PCS.COUNTRY = GEO.COUNTRY_DESCRIPTION 
    AND PCS.KEY_CUSTOMER = PC.KEY_CUSTOMER
    AND PCS.KEY_CUSTOMER = PC.KEY_CUSTOMER
    AND PC.PRODUCT_VARIANT = PCS.PRODUCT_VARIANT
    LEFT JOIN 
    (SELECT APC_CODE,TMC AS PCS_TMC1,BMC AS PCS_BMC1
    FROM [xxxx] 
    WHERE PMD_PRODUCT_STATUS = 'Active'
    **AND MILESTONE LIKE ('current_qtr_modified') -- filtered to specific qtr**
    AND NOT (PMD_PROGRAM_NAME LIKE '%CANCELED%')
    AND LOCKED IS false
    ) PCS1
    ON PC.APC_CODE = PCS1.APC_CODE
    AND PCS1.COUNTRY = GEO.COUNTRY_DESCRIPTION 
    AND PCS1.KEY_CUSTOMER = PC.KEY_CUSTOMER
    AND PCS1.PRODUCT_VARIANT = PC.PRODUCT_VARIANT

    WITH date_var AS (SELECT CURRENT_DATE current_date,EXTRACT(QUARTER FROM 
    CURRENT_DATE) As current_qtr,EXTRACT(YEAR FROM CURRENT_DATE) current_year,CONCAT(EXTRACT(YEAR FROM CURRENT_DATE),'_','Q',EXTRACT(QUARTER FROM 
    CURRENT_DATE),'M0') current_qtr_modified,CONCAT(EXTRACT(YEAR FROM 
    CURRENT_DATE),EXTRACT(QUARTER FROM CURRENT_DATE)-1,'M0') 
    prevIoUs_qtr_modified)

    SELECT current_qtr_modified,prevIoUs_qtr_modified
    FROM date_var

解决方法

您可以尝试将以下cte添加到当前查询中。

WITH
date_var
AS 
(SELECT CURRENT_DATE current_date,EXTRACT(QUARTER FROM CURRENT_DATE) As current_qtr,EXTRACT(YEAR FROM CURRENT_DATE) current_year,CONCAT(EXTRACT(YEAR FROM CURRENT_DATE),'_','Q',EXTRACT(QUARTER FROM CURRENT_DATE),'M0') current_qtr_modified,EXTRACT(QUARTER FROM CURRENT_DATE)-1,'M0') previous_qtr_modified
),-- 1st table pulling the current cost numbers based on current date from XYZ1 table

SELECT CASE WHEN PMD_PRODUCT_NAME LIKE '%MAN%' THEN 'MAN' 
    WHEN PC.COUNTRY ='BR' THEN 'Brazil'
    ELSE GEO.COUNTRY_DESCRIPTION
    END AS Country,PC.PRODUCT_STATUS AS Product_Status,PC.SA_DATE AS SA_Date,MILESTONE 
FROM
-- PC table
  (SELECT *,CASE WHEN SA_DATE < CURRENT_DATE() THEN 'Legacy' ELSE 'NPI' END PRODUCT_STATUS FROM [xxxx] 
  WHERE PMD_PRODUCT_STATUS ='Active') PC

-- Country since PC table has country codes and not description
  JOIN [xxxx] GEO
  ON PC.COUNTRY = GEO.COUNTRY_CODE
-- PCS table joined 
  LEFT JOIN 
    (SELECT APC_CODE,COUNTRY,KEY_CUSTOMER,PRODUCT_VARIANT,TMC AS PCS_TMC,BMC AS PCS_BMC,TMC_TARGET
    FROM [xxxx] 

-- 2nd table pulling the historical cost numbers (from XYZ2 table) based on MILESTONE string (**GOAL** instead basing on my query on MILESTONE string can I want it dynamic where if the current quarter changes from Q2 to Q3 and also retain the previous quarter which will be Q2)
 
    WHERE PMD_PRODUCT_STATUS = 'Active'
    **AND MILESTONE LIKE *previous_qtr_modified* -- filtered to specific qtr**
    AND NOT (PMD_PROGRAM_NAME LIKE '%CANCELED%')
    AND LOCKED IS false
    ) PCS
    ON PC.APC_CODE = PCS.APC_CODE
    AND PCS.COUNTRY = GEO.COUNTRY_DESCRIPTION 
    AND PCS.KEY_CUSTOMER = PC.KEY_CUSTOMER
    AND PCS.KEY_CUSTOMER = PC.KEY_CUSTOMER
    AND PC.PRODUCT_VARIANT = PCS.PRODUCT_VARIANT
    LEFT JOIN 
    (SELECT APC_CODE,TMC AS PCS_TMC1,BMC AS PCS_BMC1
    FROM [xxxx] 
    WHERE PMD_PRODUCT_STATUS = 'Active'
    **AND MILESTONE LIKE *current_qtr_modified* -- filtered to specific qtr**
    AND NOT (PMD_PROGRAM_NAME LIKE '%CANCELED%')
    AND LOCKED IS false
    ) PCS1
    ON PC.APC_CODE = PCS1.APC_CODE
    AND PCS1.COUNTRY = GEO.COUNTRY_DESCRIPTION 
    AND PCS1.KEY_CUSTOMER = PC.KEY_CUSTOMER
    AND PCS1.PRODUCT_VARIANT = PC.PRODUCT_VARIANT

输出：

Row  current_qtr_modified   previous_qtr_modified   
1       2020_Q3_M0              2020_Q2_M0

上述cte根据当前日期计算当前季度和上一季度，然后将年份和_Q连接起来以匹配用于过滤里程碑**AND MILESTONE LIKE ('2020_Q2_M0')的格式 使用此CTE，您只需将current_qtr_modified和previous_qtr_modified插入您当前和上一季度的硬编码位置即可。

编辑：我已经更新了查询。请参阅斜体文字。

dategoogle-bigqueryquarter