SQL Server Cumulative sum of all preceding dates in SQL

83qze16e 于 2023-03-11 发布在其他

关注(0)|答案(1)|浏览(158)

I have an inventory movement table from our ERP, and to calculate stock on hand at a specific date (X), the calculation is the sum of all Quantity where Date is less than or equal to X.

(Date is DDMMYYYY in the below example)

raw.InventoryTable -
| Date | Quantity | Item | Value |
| ------------ | ------------ | ------------ | ------------ |
| 01/01/2022 | 5 | WidgetA | 1 |
| 01/01/2023 | -5 | WidgetA | -1 |
| 31/01/2023 | 5 | WidgetA | 1 |
| 01/01/2022 | 5 | WidgetB | 5 |
| 01/02/2023 | -1 | WidgetB | -5 |
| 31/02/2023 | 1 | WidgetB | 5 |

So if I want to know inventory quantity as at end of month January 2023, it is the sum of quantity throughout all preceding time, e.g.

select 
      sum(i.quantity) as [On Hand]
    , sum(i.Value) as [Inventory Value]
    , i.Item
from raw.InventoryTable i
where Date <= '2022-01-31'
group by 
    i.item

What I want to build is an aggregate table for the end of each month (for instance just for this year), so we would have pre-agg'd stock on hand calculations and from a reporting perspective, we only have to select data from the relevant month and not worry about aggregation.

Desired output:

int.agg_Inventory_by_month
| Date | AggQuantity | Item | AggValue |
| ------------ | ------------ | ------------ | ------------ |
| 31/01/2023 | 5 | WidgetA | 1 |
| 31/01/2023 | 5 | WidgetB | 5 |
| 31/02/2023 | 5 | WidgetA | 1 |
| 31/02/2023 | 1 | WidgetB | 5 |

It's easier to perform time intelligence calculations in reporting functions with this kind of pre-agg.

I could build the aggregation in Python by holding a start date, iterating through and summing in a loop, but would ideally like to contain transformation logic exclusively in SQL (either SQL or SQL+Jinja via dbt).

sql-server

来源：https://stackoverflow.com/questions/75678416/cumulative-sum-of-all-preceding-dates-in-sql

1条答案

按热度按时间

wgeznvg71#

If I'm understanding the question correctly, it looks like you can join with dbt_utils.date_spine() to aggregate. Here's an example model using Snowflake syntax. You can swap out the SQL functions for the SQL Server equivalent and replace the start/end date with whatever you want to establish the range.

{{
  config({
    "materialized": 'table',
  })
}}

WITH DATE_RANGE AS (
  SELECT DATE_MONTH
  FROM ({{ dbt_utils.date_spine(datepart="month", start_date="TO_DATE('2022-01-01', 'YYYY-MM-DD')", end_date="TO_DATE('2022-12-31', 'YYYY-MM-DD')") }})
),
INVENTORY AS (
  SELECT * FROM {{ ref('inventory') }}
),
AGG AS (
  SELECT DR.DATE_MONTH, SUM(I.QUANTITY)
  FROM INVENTORY I
  JOIN DATE_RANGE DR ON DATE_TRUNC('MONTH', I.DATE) <= DR.DATE_MONTH
  GROUP BY 1
)

SELECT * FROM AGG

赞(0）回复(0）举报 2023-03-11

我来回答

SQL Server Cumulative sum of all preceding dates in SQL

1条答案

相关问题

热门标签

最新问答