Snowflake 中 SAP HANA 函数 workdays_between() 的等价物

Question

我有一个要求，我必须在 Snowflake 中重建 SAP HANA 视图。该视图使用 SAP HANA 中的

workdays_between()

函数，该函数用于根据工厂日历查找两个日期之间的工作日，它使用

TFACS

表来计算工作日。在 Snowflake 中我们没有

workdays_between()

功能所以请告诉我我们如何在 Snowflake 中使用这个功能来计算工作日。

我试图在 Snowflake 中创建一个 UDF，但在 Snowflake 中重新创建

workdays_between()

的逻辑非常困难，因为它根据国家代码（

TFACS

）、年份（

IDENT

）和月列。所以请帮助在 Snowflake 中创建逻辑。SAP 中的 TFACS 表： SAP 中的 TFACS 表有各种列，如 IDENT（国家代码）、JAHR（年）、MON01 ....MON12 每个月的工作日详细信息。

国家代码 TFACS 表数据：IN 年份：2023 这里 1 表示工作日，0 表示假期，包括周六和周日。

我必须在 Snowflake 中构建类似的东西。但是在 Snowflake workdays_between() 函数中不可用，如果有人有任何解决方案，请告诉我。

Answer 1

JAHR

例如：

create or replace function workdays_between(
    date1 date, date2 date, country_region string
)
returns number
language python
runtime_version = '3.8'
packages = ('holidays')
handler='x'
as 
$$
from datetime import date, timedelta
import holidays
def x(start_date, end_date, country_region):
    days = (end_date - start_date).days + 1
    if days <= 0:
        return 0

    parsed_country_region = country_region.split('-')
    country = parsed_country_region[0]
    region = (
        parsed_country_region[1]
        if len(parsed_country_region) > 1
        else None
    )
    range_years = range(start_date.year, end_date.year+1)
    holidays_dict = holidays.country_holidays(
       country
       , subdiv=region
       , years=range_years
    )

    weekdays = [0, 1, 2, 3, 4]

    working_days = [day for day in range(days) if (start_date + timedelta(day)).weekday() in weekdays and
        (start_date + timedelta(day)) not in holidays_dict]
    return len(working_days)
$$
;

相关：

https://medium.com/snowflake/generating-all-the-holidays-in-sql-with-a-python-udtf-4397f190252b

Answer 2

select workdays_between('2022-01-01', '2022-10-10', 'US-CA');

-- 192

功能。

完整的文章可以在

这里

找到，但是解决方案代码是这样的： workdays_between

现在我可以简单地跑了

create view SF_SAP_FACTORY_CALENDAR as (
    with tfacs_pivot as (
        select IDENT, JAHR, MON01 as DAY_FLAGS, '01' as "MONTH" from TFACS UNION ALL 
        select IDENT, JAHR, MON02 as DAY_FLAGS, '02' as "MONTH" from TFACS UNION ALL 
        select IDENT, JAHR, MON03 as DAY_FLAGS, '03' as "MONTH" from TFACS UNION ALL 
        select IDENT, JAHR, MON04 as DAY_FLAGS, '04' as "MONTH" from TFACS UNION ALL 
        select IDENT, JAHR, MON05 as DAY_FLAGS, '05' as "MONTH" from TFACS UNION ALL 
        select IDENT, JAHR, MON06 as DAY_FLAGS, '06' as "MONTH" from TFACS UNION ALL 
        select IDENT, JAHR, MON07 as DAY_FLAGS, '07' as "MONTH" from TFACS UNION ALL 
        select IDENT, JAHR, MON08 as DAY_FLAGS, '08' as "MONTH" from TFACS UNION ALL 
        select IDENT, JAHR, MON09 as DAY_FLAGS, '09' as "MONTH" from TFACS UNION ALL 
        select IDENT, JAHR, MON10 as DAY_FLAGS, '10' as "MONTH" from TFACS UNION ALL 
        select IDENT, JAHR, MON11 as DAY_FLAGS, '11' as "MONTH" from TFACS UNION ALL 
        select IDENT, JAHR, MON12 as DAY_FLAGS, '12' as "MONTH" from TFACS),
     day_point as (
        select tfp.*
                , length(DAY_FLAGS) AS NUM_FLAGGED_DAYS
                ,  ROW_NUMBER() OVER (PARTITION BY "JAHR", "MONTH" ORDER BY "JAHR", "MONTH" )  as FLAG_POINTER
            from tfacs_pivot tfp
                cross join table ( generator (rowcount =>32))  d
     
     )
    select dp.*
            , date_from_parts (JAHR, "MONTH", FLAG_POINTER) as FACTORY_DATE
            , substr(dp.DAY_FLAGS, FLAG_POINTER , 1) as day_marker
            , to_boolean (DAY_MARKER) as IS_WORKING_DAY
        from day_point dp
    where FLAG_POINTER <= NUM_FLAGGED_DAYS);

获取我想要的信息：

select FACTORY_DATE, IS_WORKING_DAY from SF_SAP_FACTORY_CALENDAR;

设置此视图并可能在生产环境中具体化后，创建模仿 SAP HANA 等价物的 

FACTORY_DATE   IS_WORKING_DAY
2018-01-01     FALSE
2018-01-02     TRUE
2018-01-03     TRUE
2018-01-04     TRUE
2018-01-05     TRUE
2018-01-06     FALSE
2018-01-07     FALSE
2018-01-08     TRUE
...            ...

功能只是一个小的附加步骤：

WORKDAYS_BETWEEN