update20260427

2026-04-27 11:04:09 +08:00
commit 3e14e78af3
173 changed files with 62579 additions and 0 deletions
--- a/tmp_ims_tf_fact_sales.sql
+++ b/tmp_ims_tf_fact_sales.sql
@@ -0,0 +1,317 @@
+-- Databricks notebook source
+-- 20260130 临时方案，解决测试环境有NULL值问题
+create or replace temporary view dwd_ims_td_pack
+as 
+select 
+  distinct
+  pack_id,
+  pack_code
+from dwd.dwd_ims_td_pack
+
+-- COMMAND ----------
+
+CREATE OR REPLACE TEMPORARY VIEW FACT_CHPA_SALES_TEMP_WITH_PREVIOUS
+AS
+SELECT
+  CAST(YM AS INT) AS YM,
+  CAST(YEAR AS INT) AS YEAR,
+  CAST(REPLACE(YM,YEAR,'') AS INT) AS  MONTH,
+  CAST(VALUE AS DECIMAL(38,10)) AS VALUE,
+  CAST(COUNTINGUNIT AS DECIMAL(38,10)) AS COUNTINGUNIT,
+  CAST(TOTALUNIT AS DECIMAL(38,10)) AS TOTALUNIT,
+  PROVINCE_C,
+  PHCD,
+  conversion_ratio  
+FROM dwd.dwd_gnd_pharbers_prov_fact
+
+-- COMMAND ----------
+
+CREATE OR REPLACE TEMPORARY VIEW DIM_PROVINCE
+AS
+select 
+  -- t2.geo_key , 
+  t2.province_code,
+  CASE 
+    WHEN 
+      t2.province_name 
+        IN ('上海','北京','天津','重庆') 
+          THEN
+            CONCAT(t2.province_name,'市')
+    WHEN 
+      t2.province_name 
+        IN ('内蒙古','西藏') 
+          THEN
+            CONCAT(t2.province_name,'自治区')      
+    WHEN 
+      t2.province_name = '宁夏'
+        THEN CONCAT(t2.province_name,'回族自治区')  
+    WHEN 
+      t2.province_name = '广西'
+        THEN CONCAT(t2.province_name,'壮族自治区') 
+    WHEN 
+      t2.province_name = '新疆'
+        THEN CONCAT(t2.province_name,'维吾尔自治区')   
+          ELSE CONCAT(t2.province_name,'省')     
+            END AS province_name            
+from dm.dm_td_geography t2 
+where t2.geo_key = t2.province_code
+and province_name not in ('台湾','澳门','香港')
+
+-- COMMAND ----------
+
+CREATE OR REPLACE TEMPORARY VIEW DIM_CHPA_PACK_INFO
+AS
+SELECT
+  PACK_COD,
+  MAX(IQVIA_PACK_CODE) AS IQVIA_PACK_CODE,
+  MAX(countingunit) as COUNTINGUNIT_RATIO
+FROM
+  dwd.dwd_gnd_dept_pack_property
+GROUP BY
+  PACK_COD
+
+-- COMMAND ----------
+
+CREATE OR REPLACE TEMPORARY VIEW FACT_CHPA_SALES
+AS 
+SELECT 
+	YM,
+	YEAR,
+	REPLACE(YM,YEAR,'') AS  MONTH,
+  T3.IQVIA_PACK_CODE AS PACK_CODE,
+	T2.PROVINCE_CODE AS AUDIT_COD,
+	VALUE AS MTH00LC,
+	0 AS MTH00LCLY,
+	CASE WHEN T3.COUNTINGUNIT_RATIO IS NULL THEN TOTALUNIT*T1.conversion_ratio ELSE TOTALUNIT*T3.COUNTINGUNIT_RATIO END AS MTH00CN,
+	0 AS MTH00CNLY,
+	TOTALUNIT AS MTH00UN,
+	0 AS MTH00UNLY
+FROM FACT_CHPA_SALES_TEMP_WITH_PREVIOUS AS T1
+LEFT JOIN DIM_PROVINCE AS T2 ON T1.PROVINCE_C = T2.PROVINCE_NAME
+LEFT JOIN DIM_CHPA_PACK_INFO AS T3 ON T1.PHCD = T3.PACK_COD
+UNION ALL
+SELECT 
+	CAST((YEAR+1)*100+REPLACE(YM,YEAR,'') AS INT) AS YM,
+	YEAR+1,
+	CAST(REPLACE(YM,YEAR,'') AS INT) AS  MONTH,
+  T3.IQVIA_PACK_CODE AS PACK_CODE,
+	T2.PROVINCE_CODE AS AUDIT_COD,
+	0 AS MTH00LC,
+	VALUE AS MTH00LCLY,
+	0 AS MTH00CN,
+	CASE WHEN T3.COUNTINGUNIT_RATIO IS NULL THEN TOTALUNIT*T1.conversion_ratio ELSE TOTALUNIT*T3.COUNTINGUNIT_RATIO END AS MTH00CNLY,
+	0 AS MTH00UN,
+	TOTALUNIT AS MTH00UNLY
+FROM FACT_CHPA_SALES_TEMP_WITH_PREVIOUS AS T1
+LEFT JOIN DIM_PROVINCE AS T2 ON T1.PROVINCE_C = T2.PROVINCE_NAME
+LEFT JOIN DIM_CHPA_PACK_INFO AS T3 ON T1.PHCD = T3.PACK_COD
+
+-- COMMAND ----------
+
+-- 20260119 dwd.dwd_ims_td_pack、dwd.dwd_ims_tf_fact_sales 不依赖
+--删除最近五年的dim pack数据，将最新的dim pack写入
+delete from dws.dws_ims_td_pack_ym
+where ym + 500 > 
+  (select max(year * 100 + month) from dwd.dwd_ims_tf_fact_sales);
+insert into dws.dws_ims_td_pack_ym
+select
+  distinct ym,
+  pack_id,
+  pack_code,
+  from_utc_timestamp(current_timestamp(), 'UTC+8'),
+  from_utc_timestamp(current_timestamp(), 'UTC+8')
+from
+  (
+    select
+      distinct year * 100 + month as ym
+    from
+      dwd.dwd_ims_tf_fact_sales
+    where 
+      year * 100 + month + 500 > (
+        select max(year * 100 + month) from dwd.dwd_ims_tf_fact_sales
+      )
+  ) full
+  -- join dwd.dwd_ims_td_pack
+  join dwd_ims_td_pack
+  ;
+
+-- COMMAND ----------
+
+-- --IMS 原始数据预处理
+-- --FACT SALES 处理
+-- insert overwrite table tmp.tmp_ims_tf_fact_sales 
+-- select distinct
+-- 	FACT.YM,AUDIT.Audit_Code AS AUDIT_COD,FACT.Pack_Code AS PACK_COD,
+-- 	sum(MTH00LC) as MTH00LC,sum(MTH00LCLY) as MTH00LCLY,
+-- 	sum(MTH00CN) as MTH00CN,sum(MTH00CNLY) as MTH00CNLY,
+-- 	sum(MTH00UN) as MTH00UN,sum(MTH00UNLY) as MTH00UNLY
+-- from(
+--   select 
+-- 	Year*100+Month as YM,
+-- 	Year ,
+-- 	Month ,
+-- 	t2.pack_code ,
+-- 	Audit_ID ,
+-- 	SalesValue_LC as MTH00LC,
+-- 	0 as MTH00LCLY,
+-- 	Counting_Unit as MTH00CN,
+-- 	0 as MTH00CNLY,
+-- 	Sales_Unit as MTH00UN,
+-- 	0 as MTH00UNLY
+-- 	from dwd.dwd_ims_tf_fact_sales t1
+-- --补充历史的PACK ID 和 PACK CODE的关联关系，带上YM将最新dim pack表full join 最近五年
+--   left join dws.dws_ims_td_pack_ym t2 on t1.Pack_ID=t2.pack_id and t1.Year*100+t1.Month=t2.ym
+-- 	union all
+-- 	select 
+-- 	(Year+1)*100+Month as YM,
+-- 	Year+1 ,
+-- 	Month ,
+-- 	t2.pack_code ,
+-- 	Audit_ID ,
+-- 	0 as MTH00LC,
+-- 	SalesValue_LC as MTH00LCLY,
+-- 	0 as MTH00CN,
+-- 	Counting_Unit as MTH00CNLY,
+-- 	0 as MTH00UN,
+-- 	Sales_Unit as MTH00UNLY
+-- 	from dwd.dwd_ims_tf_fact_sales t1
+-- --补充历史的PACK ID 和 PACK CODE的关联关系，带上YM将最新dim pack表full join 最近五年
+--   left join dws.dws_ims_td_pack_ym t2 on t1.Pack_ID=t2.pack_id and t1.Year*100+t1.Month=t2.ym
+-- )FACT 
+-- LEFT JOIN dwd.dwd_ims_td_audit AUDIT
+-- ON FACT.Audit_ID = AUDIT.Audit_ID
+-- where FACT.YM<=(select max(Year*100+Month) from dwd.dwd_ims_tf_fact_sales)
+-- ----------20260320 chenwu CHPA的数据只保留202201之后的
+-- and FACT.YM>=202201
+-- ----------
+-- AND AUDIT.Audit_Code = 'CHT'
+-- group by FACT.YM,AUDIT.Audit_Code,FACT.Pack_Code 
+
+
+
+-- -- 20260120新增
+-- -- 以前 全国和省份数据放在一起，现在全国是上面，省份是新的，单独放在下面。
+-- union all
+-- select distinct
+-- 	CAST(YM AS INT) AS YM,
+-- 	AUDIT_COD,Pack_Code AS PACK_COD,
+-- 	sum(MTH00LC) as MTH00LC,sum(MTH00LCLY) as MTH00LCLY,
+-- 	sum(MTH00CN) as MTH00CN,sum(MTH00CNLY) as MTH00CNLY,
+-- 	sum(MTH00UN) as MTH00UN,sum(MTH00UNLY) as MTH00UNLY
+-- from	
+-- 	FACT_CHPA_SALES
+-- where 
+--   YM<=(select max(YM) from dwd.dwd_gnd_pharbers_prov_fact)
+-- group by YM,Audit_Cod,Pack_Code 
+-- ;
+
+
+-- COMMAND ----------
+
+-- IMS 原始数据预处理 —— FACT SALES 处理
+INSERT OVERWRITE TABLE tmp.tmp_ims_tf_fact_sales
+
+WITH yearmont_range (--防止翘角导致数据存在差异
+  select
+  min(ym) YM
+from
+  (
+    select
+      max(ym) ym
+    from
+      dwd.dwd_gnd_pharbers_prov_fact
+    union all
+    select
+      max(Year * 100 + Month) ym
+    from
+      dwd.dwd_ims_tf_fact_sales
+  )
+)
+
+-- ════════════════════════════════════════════
+-- Part 1：全国数据（CHT）
+-- ════════════════════════════════════════════
+SELECT DISTINCT
+    FACT.YM,
+    AUDIT.Audit_Code  AS AUDIT_COD,
+    FACT.Pack_Code    AS PACK_COD,
+    SUM(MTH00LC)      AS MTH00LC,
+    SUM(MTH00LCLY)    AS MTH00LCLY,
+    SUM(MTH00CN)      AS MTH00CN,
+    SUM(MTH00CNLY)    AS MTH00CNLY,
+    SUM(MTH00UN)      AS MTH00UN,
+    SUM(MTH00UNLY)    AS MTH00UNLY
+FROM (
+    -- 当期销售数据
+    SELECT
+        Year * 100 + Month  AS YM,
+        Year,
+        Month,
+        t2.pack_code,
+        Audit_ID,
+        SalesValue_LC       AS MTH00LC,
+        0                   AS MTH00LCLY,
+        Counting_Unit       AS MTH00CN,
+        0                   AS MTH00CNLY,
+        Sales_Unit          AS MTH00UN,
+        0                   AS MTH00UNLY
+    FROM dwd.dwd_ims_tf_fact_sales  t1
+    -- 补充历史 Pack_ID → Pack_Code 映射（带 YM，关联最近五年快照）
+    LEFT JOIN dws.dws_ims_td_pack_ym t2
+           ON t1.Pack_ID = t2.pack_id
+          AND t1.Year * 100 + t1.Month = t2.ym
+
+    UNION ALL
+
+    -- 同期数据（当期行 YM+1 年，作为去年同期值）
+    SELECT
+        (Year + 1) * 100 + Month  AS YM,
+        Year + 1,
+        Month,
+        t2.pack_code,
+        Audit_ID,
+        0                         AS MTH00LC,
+        SalesValue_LC             AS MTH00LCLY,
+        0                         AS MTH00CN,
+        Counting_Unit             AS MTH00CNLY,
+        0                         AS MTH00UN,
+        Sales_Unit                AS MTH00UNLY
+    FROM dwd.dwd_ims_tf_fact_sales   t1
+    -- 补充历史 Pack_ID → Pack_Code 映射（带 YM，关联最近五年快照）
+    LEFT JOIN dws.dws_ims_td_pack_ym t2
+           ON t1.Pack_ID = t2.pack_id
+          AND t1.Year * 100 + t1.Month = t2.ym
+) FACT
+LEFT JOIN dwd.dwd_ims_td_audit AUDIT
+       ON FACT.Audit_ID = AUDIT.Audit_ID
+WHERE FACT.YM <= (SELECT YM FROM yearmont_range)
+  AND FACT.YM >= 202201          -- 20260320 chenwu CHPA 数据只保留 202201 之后
+  AND AUDIT.Audit_Code = 'CHT'
+GROUP BY
+    FACT.YM,
+    AUDIT.Audit_Code,
+    FACT.Pack_Code
+
+UNION ALL
+
+-- ════════════════════════════════════════════
+-- Part 2：省份数据（CHPA）
+-- 20260120 新增：全国与省份拆分，省份数据单独维护
+-- ════════════════════════════════════════════
+SELECT DISTINCT
+    CAST(YM AS INT)  AS YM,
+    AUDIT_COD,
+    Pack_Code        AS PACK_COD,
+    SUM(MTH00LC)     AS MTH00LC,
+    SUM(MTH00LCLY)   AS MTH00LCLY,
+    SUM(MTH00CN)     AS MTH00CN,
+    SUM(MTH00CNLY)   AS MTH00CNLY,
+    SUM(MTH00UN)     AS MTH00UN,
+    SUM(MTH00UNLY)   AS MTH00UNLY
+FROM FACT_CHPA_SALES
+WHERE YM <= (SELECT YM FROM yearmont_range)
+GROUP BY
+    YM,
+    Audit_Cod,
+    Pack_Code
+;