How to calculate age based on current date - sql

Question

I am trying to calculate the age of an individual.
I would like the age to be calculated to the exact as possible. However, if the age is on the same date, it should calculate the age assume it's already his birthday.
EXAMPLE: Today date: Dec 15, 2020
user #1: DOB: Dec 14, 2005 (AGE: 15)
user #2: DOB: Dec 16, 2005 (AGE: 14)
user #3: DOB: Dec 15, 2005 (AGE: 15)
I have this query, but for user #3, it is giving me age 14 when it should be 15.
SELECT
USER,
DATE_OF_BIRTH,
FLOOR(DATEDIFF(DAY, B.DATE_OF_BIRTH, GETDATE())/365.25) AS CALCULATED_AGE
FROM TBL_A

Answer 1

I think you could do it using the MONTHS_BETWEEN function.
select floor(months_between(getdate(), date_of_birth) / 12) from tbl_a

Answer 2

I have to assume this question is not being done in Snowflake, as the results are off by one.
with mytable(user, dob, age_expected) as (
select * from values
(1, '2005-12-14'::date, 15),
(2, '2005-12-16'::date, 14),
(3, '2005-12-15'::date, 15)
)
SELECT
user
,dob
,'2022-12-15'::date as todays_date
,DATEDIFF(DAY, dob, todays_date) as gap_days
,gap_days/365.25 as d365
,floor(d365) as f0
,FLOOR(DATEDIFF(DAY, dob, todays_date)/365.25) AS CALCULATED_AGE
FROM mytable
USER
DOB
TODAYS_DATE
GAP_DAYS
D365
F0
CALCULATED_AGE
1
2005-12-14
2022-12-15
6,210
17.002053
17
17
2
2005-12-16
2022-12-15
6,208
16.996578
16
16
3
2005-12-15
2022-12-15
6,209
16.999316
16
16
So if we deconstruct what is happening, move the month of interest back to november so the before after months case can be seen, and move the "current date" back 2 years so we get the stated expected results 14/15:
with mytable(user, dob, age_expected) as (
select * from values
(1, '2005-10-15'::date, 15),
(2, '2005-11-14'::date, 15),
(3, '2005-11-16'::date, 14),
(4, '2005-11-15'::date, 15),
(5, '2005-12-15'::date, 14)
)
SELECT
user
,dob
,age_expected
,'2020-11-15'::date as todays_date
,year(dob) as dob_y
,month(dob) as dob_m
,day(dob) as dob_d
,year(todays_date) as today_y
,month(todays_date) as today_m
,day(todays_date) as today_d
,DATEDIFF(DAY, dob, todays_date)/365.25 as float_val
,today_y - dob_y as y_dif
,case
when today_m < dob_m then 'a'
when today_m > dob_m then 'b'
when today_d < dob_d then 'c'
when today_d >= dob_d then 'd'
end as pick
,case
when today_m < dob_m then y_dif -1
when today_m > dob_m then y_dif
when today_d < dob_d then y_dif -1
when today_d >= dob_d then y_dif
end as result
FROM mytable
order by dob
USER
DOB
AGE_EXPECTED
TODAYS_DATE
DOB_Y
DOB_M
DOB_D
TODAY_Y
TODAY_M
TODAY_D
FLOAT_VAL
Y_DIF
PICK
RESULT
1
2005-10-15
15
2020-11-15
2,005
10
15
2,020
11
15
15.085558
15
b
15
2
2005-11-14
15
2020-11-15
2,005
11
14
2,020
11
15
15.003422
15
d
15
4
2005-11-15
15
2020-11-15
2,005
11
15
2,020
11
15
15.000684
15
d
15
3
2005-11-16
14
2020-11-15
2,005
11
16
2,020
11
15
14.997947
15
c
14
5
2005-12-15
14
2020-11-15
2,005
12
15
2,020
11
15
14.918549
15
a
14
So using the numbers you state, would would, but it we move the current date back to 2022 those numbers don't work, due to the stride of actual number of daylight days. so you just stick to the partial math method which can be simplefied to:
with mytable(user, dob, age_expected) as (
select * from values
(1, '2005-11-15'::date, 15),
(2, '2005-12-14'::date, 15),
(3, '2005-12-16'::date, 14),
(4, '2005-12-15'::date, 15),
(5, '2006-01-15'::date, 14)
)
SELECT
user
,dob
,age_expected
,CURRENT_DATE
,case
when month(CURRENT_DATE) < month(dob) then year(CURRENT_DATE) - year(dob) -1
when month(CURRENT_DATE) > month(dob) then year(CURRENT_DATE) - year(dob)
when day(CURRENT_DATE) < day(dob) then year(CURRENT_DATE) - year(dob) -1
when day(CURRENT_DATE) >= day(dob) then year(CURRENT_DATE) - year(dob)
end as result
FROM mytable
order by dob
gives:
USER
DOB
AGE_EXPECTED
CURRENT_DATE
RESULT
1
2005-11-15
15
2022-12-15
17
2
2005-12-14
15
2022-12-15
17
4
2005-12-15
15
2022-12-15
17
3
2005-12-16
14
2022-12-15
16
5
2006-01-15
14
2022-12-15
16

Answer 3

Here is an overloaded UDF you can use. The one that takes a single parameter says you want a person's age as of today. The one that takes two parameters says you want a person's age as of the date specified in the second parameter.
create or replace function age(DATE_OF_BIRTH date)
returns integer
language sql
as
$$
year(current_date) - year(DATE_OF_BIRTH) - iff(month(current_date) <= month(DATE_OF_BIRTH) and day(current_date) < day(DATE_OF_BIRTH), 1, 0)
$$;
create or replace function age(DATE_OF_BIRTH date, AS_OF_DATE date)
returns integer
language sql
as
$$
year(AS_OF_DATE) - year(DATE_OF_BIRTH) - iff(month(AS_OF_DATE) <= month(DATE_OF_BIRTH) and day(AS_OF_DATE) < day(DATE_OF_BIRTH), 1, 0)
$$;
Tests with data set for the current date:
with PEOPLE as
(
select COLUMN1::date DOB from values
('2000-01-01'),
('2000-01-02'),
('2000-12-14'),
('2000-12-15'),
('2000-12-16'),
('2000-12-17')
)
select age(DOB) from PEOPLE
;
AGE(DOB)
22
22
22
22
22
21
This test is using an as of date to check people's ages:
with PEOPLE as
(
select COLUMN1::date DOB from values
('2000-01-01'),
('2000-01-02'),
('2000-12-14'),
('2000-12-15'),
('2000-12-16'),
('2000-12-17')
)
select age(DOB, '2023-01-01') from PEOPLE
;
AGE(DOB, '2023-01-01')
23
22
22
22
22
22