How to add an integer to a date field in Snowflake? - sql

Question

Have a WHERE clause in my SQL query in Snowflake where I would like to filter by adding an integer to a date field:
AND CAST(DATETIME AS DATE) >= '2022-10-26' + 30
SNK returns me this error: invalid type [TO_DATE((2022 - 10) - 26)] for parameter 'TO_DATE'
Anyone know how to overturn this error?

Answer 1

Use the DATEADD() function to add days:
CAST(DATETIME AS DATE) >= DATEADD(day, 30, '2022-10-26')

Answer 2

Using INTERVAL arithmetic:
CAST(DATETIME AS DATE) >= '2022-10-26'::DATE + INTERVAL '30 DAYS'

Answer 3

the base problem '2022-10-26' + 30 is that a string and number do not implicitly compare to a date
So there are a couple of ways to do the cast:
select
column1 as datetime
,to_date(datetime, 'yyyy-mm-dd') as d1
,try_to_date(datetime, 'yyyy-mm-dd') as d2
,datetime::date as d3
,cast(datetime as date) as d4
--,d1 + 30 as add1
-- ,dateadd(day, 30, '2022-10-26') as add2
--,d1 + interval '30 days'
from values
('2023-02-10'),
('2022-10-26');
DATETIME
D1
D2
D3
D4
2023-02-10
2023-02-10
2023-02-10
2023-02-10
2023-02-10
2022-10-26
2022-10-26
2022-10-26
2022-10-26
2022-10-26
and the second half is addition
select
column1 as datetime
,to_date(datetime, 'yyyy-mm-dd') as d1
,d1 + 30 as add1
,dateadd(day, 30, '2022-10-26') as add2
,d1 + interval '30 days' as add3
from values
('2023-02-10'),
('2022-10-26')
DATETIME
D1
ADD1
ADD2
ADD3
2023-02-10
2023-02-10
2023-03-12
2022-11-25 00:00:00.000
2023-03-12
2022-10-26
2022-10-26
2022-11-25
2022-11-25 00:00:00.000
2022-11-25
this shows that DATEADD can handle being given a string (constant of dynamic) and the compiler will correctly cast the string to DATE, but then the DATEADD givens you a TIMESTAMP back. But the two inline addition methods require the date side of the addition to already be a date, thus it needs casting in one fashion or another.