Find the last weekday for a given month in PostgreSQL
Usage: If month end falls on a Saturday or a Sunday, return the previous Friday, else use month end
Examples:
3/31/2013 falls on a Sunday, so return 3/29/2013
11/30/2013 falls on a Saturday, so return 11/29/2013
How to write this in PostgreSQL SQL?
What I have so far is this (returns only Month Ends, but month ends don't exist when they fall on a Saturday or Sunday):
SELECT as_of_dt, sum(bank_shr_bal) as bank_shr_bal
FROM hm_101.vw_gl_bal
WHERE as_of_dt = (date_trunc('MONTH', as_of_dt) + INTERVAL '1 MONTH - 1 day')::date
GROUP BY 1
Thanks
A calendar table really simplifies the SQL for queries like these. (The table "weekdays" is actually a view based on the calendar table. The structure of it should be obvious.)
select max(cal_date)
from weekdays
where cal_date < '2013-05-01'
or
select max(cal_date)
from weekdays
where cal_date between '2013-04-01' and '2013-04-30'
with s as (
select *, (date_trunc('MONTH', as_of_dt) + INTERVAL '1 MONTH - 1 day')::date last_day
from
hm_101.vw_gl_bal
)
SELECT
as_of_dt,
gl_acct_nbr,
cc_nbr,
sum(bank_shr_bal) as bank_shr_bal
FROM s
WHERE as_of_dt = (
last_day
-
(extract(dow from last_day) = 5)::int
-
2 * (extract(dow from last_day) = 6)::int
)
GROUP BY 1,2,3
One solution is to use a CTE, find the last day by month in the data and the actual last day for each month
WITH s1
as
(
SELECT
date_part('YEAR', as_of_dt) AOD_Year
,date_part('MONTH', as_of_dt) AOD_Month
,(date_trunc('MONTH', as_of_dt) + INTERVAL '1 MONTH - 1 day')::date AOD_MonthEnd
,max(as_of_dt) AOD_LastFound
FROM hm_101.vw_gl_bal
where (date_trunc('MONTH', as_of_dt) + INTERVAL '1 MONTH - 1 day')::date = '2013-03-31'
group by 1, 2, 3
)
SELECT
s1.AOD_MonthEnd
,s1.AOD_LastFound
,sum(v.bank_shr_bal) as bank_shr_bal
FROM hm_101.vw_gl_bal v
INNER JOIN s1
on v.as_of_dt = s1.AOD_LastFound
WHERE v.as_of_dt = '2013-03-29'
GROUP BY 1, 2
What you want to do is remove between 0 and 2 days from the last day of the month (which you have).
By extracting the Day Of Week (DOW) and check if it's 0 (Sunday) or 6 (Saturday), we know how many days to remove.
You can do it like this:
... - INTERVAL '1 day' * CASE date_part('DOW', last_day_of_month)
WHEN 0 THEN 2 -- Sunday, remove 2 days.
WHEN 6 THEN 1 -- Saturday, remove 1 day.
ELSE 0 -- Don't remove any days.
END
For the sake of readability I didn't include the complete last_day_of_month-calculation in there.
You can actually do this without CTEs or stored procedures.
select
case
when extract(dow from last_day_of_month) = 0
then last_day_of_month - 2
when extract(dow from last_day_of_month) = 6
then last_day_of_month - 1
else
last_day_of_month
end as last_weekday_of_month
from(
SELECT (date_trunc('MONTH', as_of_dt)
+ INTERVAL '1 MONTH - 1 day')::date as last_day_of_month
from hm_101.vw_gl_bal
)subquery;
select
case
when extract(dow from first_day_of_month) = 0 then first_day_of_month
when extract(dow from first_day_of_month) = 1 then first_day_of_month - 1
when extract(dow from first_day_of_month) = 2 then first_day_of_month - 2
when extract(dow from first_day_of_month) = 3 then first_day_of_month - 3
when extract(dow from first_day_of_month) = 4 then first_day_of_month - 4
when extract(dow from first_day_of_month) = 5 then first_day_of_month - 5
when extract(dow from first_day_of_month) = 6 then first_day_of_month - 6
end as first_weekday_of_month,
case
when extract(dow from last_day_of_month) = 6 then last_day_of_month
when extract(dow from last_day_of_month) = 5 then last_day_of_month - 6
when extract(dow from last_day_of_month) = 4 then last_day_of_month - 5
when extract(dow from last_day_of_month) = 3 then last_day_of_month - 4
when extract(dow from last_day_of_month) = 2 then last_day_of_month - 3
when extract(dow from last_day_of_month) = 1 then last_day_of_month - 2
when extract(dow from last_day_of_month) = 0 then last_day_of_month - 1
end as last_weekday_of_month
from(
SELECT
(date_trunc('month', current_date) -'7day'::interval)::date first_day_of_month,
(date_trunc('month', current_date) -'1day'::interval)::date as last_day_of_month
)subquery;
Related
I have created the following calendar table:
WITH dates AS (
SELECT EXPLODE(SEQUENCE(TO_DATE('1970-01-01'), TO_DATE('2100-12-31'), INTERVAL 1 DAY)) AS calendar_date
),
calendar_table AS (
SELECT
YEAR(calendar_date) * 10000 + MONTH(calendar_date) * 100 + DAY(calendar_date) AS date_integer,
calendar_date,
YEAR(calendar_date) AS year_of_date,
QUARTER(calendar_date) AS quarter_of_year,
MONTH(calendar_date) AS month_of_year,
DAY(calendar_date) AS day_of_month,
WEEKDAY(calendar_date) + 1 AS day_of_week_start_monday,
DAYOFWEEK(calendar_date) AS day_of_week_start_sunday,
CASE
WHEN DAY(calendar_date) >= 1 AND DAY(calendar_date) <= 7 THEN 1
WHEN DAY(calendar_date) >= 8 AND DAY(calendar_date) <= 14 THEN 2
WHEN DAY(calendar_date) >= 15 AND DAY(calendar_date) <= 21 THEN 3
WHEN DAY(calendar_date) >= 22 AND DAY(calendar_date) <= 28 THEN 4
ELSE 5
END AS day_of_week_ordinal,
CASE
WHEN WEEKDAY(calendar_date) < 5 THEN TRUE
ELSE FALSE
END AS is_week_day,
CASE
WHEN WEEKDAY(calendar_date) > 4 THEN TRUE
ELSE FALSE
END AS is_weekend,
CASE
WHEN calendar_date = DATE_TRUNC('month', calendar_date)::DATE THEN TRUE
ELSE FALSE
END AS is_first_day_of_month,
CASE
WHEN calendar_date = LAST_DAY(calendar_date) THEN TRUE
ELSE FALSE
END AS is_last_day_of_month,
DAYOFYEAR(calendar_date) AS day_of_year,
WEEKOFYEAR(calendar_date) AS iso_week_of_year,
EXTRACT(YEAROFWEEK FROM calendar_date) AS iso_year_of_date,
FROM
dates
)
I am missing a custom calendar column that would abide by the following rule:
From the second Sunday (inclusive) in December of each year, the column should contain a concatenation of 'X' and the year number of the following year.
Example:
calendar_date
custom_column
2022-12-10
X2022
2022-12-11
X2023
2022-12-12
X2023
...
...
2023-12-09
X2023
2023-12-10
X2024
2023-12-11
X2024
So far, I've been able to identify the second Sunday of December in each year by combining the logic behind the columns month_of_year, day_of_week_ordinal and day_of_week_start_monday in my calendar table, but I fail to grasp any implementation (I'm sure I'm missing something simple here).
I can calculate a flag for the second Sunday in December of each year by utilising the following logic:
CASE
WHEN
month_of_year = 12
AND day_of_week_ordinal = 2
AND day_of_week_start_monday = 7 THEN TRUE
ELSE FALSE
END AS second_sunday_in_month
But I fail to see how I can get transfer this logic to what I want as the end result.
Edit: I have added a PostgreSQL fiddle as an interactive example.
Here this might be able to help you :
SELECT
*,
CASE
WHEN calendar_date between (select calendar_date from calendar_table
where month_of_year = 12 -- December
AND day_of_week_start_monday = 7 -- Sunday
AND day_of_week_ordinal = 2 ) and (select date_trunc('year',calendar_date + interval '1 year') - interval '1 day')
THEN 'X' || date_part('year', calendar_date) +1
ELSE 'X' || date_part('year', calendar_date)
END AS is_second_sunday_of_december
FROM calendar_table;
https://dbfiddle.uk/WynGf5w_
Problem is its only working on yearly basis so there might be more tweaking needed!
Update:
There you go:
CASE
WHEN
year_of_date = (select year_of_date from calendar_table where month_of_year = 12 -- December
AND day_of_week_start_monday = 7 -- Sunday
AND day_of_week_ordinal = 2)
AND calendar_date between (select calendar_date from calendar_table
where month_of_year = 12 -- December
AND day_of_week_start_monday = 7 -- Sunday
AND day_of_week_ordinal = 2 ) and (select date_trunc('year',calendar_date + interval '1 year') - interval '1 day')
THEN 'X' || date_part('year', calendar_date) + 1
else 'X' || date_part('year', calendar_date)
END AS is_second_sunday_of_december
from calendar_table
fiddle
i hope this time the link works if not there is an underscore on the end.
I am trying to create an SQL UDF in bigQuery to calculate week in month. I got the result that I'm expecting but my function looks super messy.
create or replace function internal.week_in_month(my_date TIMESTAMP)
returns FLOAT64 as
(
case when
(case when EXTRACT(DAYOFWEEK FROM date_add(date(my_date), INTERVAL -(EXTRACT(DAY FROM my_date))+1 day)) = 1 then 7
else EXTRACT(DAYOFWEEK FROM date_add(date(my_date), INTERVAL -(EXTRACT(DAY FROM my_date))+1 day)) -1 end) > 1 then -- check first day of month to decide if it's a complete week (starts on Monday)
case when EXTRACT(DAY FROM my_date) <= 7 then -- for incomplete week
case when
(case when EXTRACT(DAYOFWEEK FROM my_date) = 1 then 7 else EXTRACT(DAYOFWEEK FROM my_date)-1 end) - EXTRACT(DAY FROM my_date) =
(case when EXTRACT(DAYOFWEEK FROM date_add(date(my_date), INTERVAL -(EXTRACT(DAY FROM my_date))+1 day)) = 1 then 7
else EXTRACT(DAYOFWEEK FROM date_add(date(my_date), INTERVAL -(EXTRACT(DAY FROM my_date))+1 day)) -1 end) -1 then 1 -- incomplete week 1
else FLOOR(( EXTRACT(DAY FROM my_date) + (case when EXTRACT(DAYOFWEEK FROM date_add(date(my_date), INTERVAL -(EXTRACT(DAY FROM my_date))+1 day)) = 1 then 7
else EXTRACT(DAYOFWEEK FROM date_add(date(my_date), INTERVAL -(EXTRACT(DAY FROM my_date))+1 day)) -1 end) -2 )/7)+1 end -- calculate week based on date
else FLOOR(( EXTRACT(DAY FROM my_date) + (case when EXTRACT(DAYOFWEEK FROM date_add(date(my_date), INTERVAL -(EXTRACT(DAY FROM my_date))+1 day)) = 1 then 7
else EXTRACT(DAYOFWEEK FROM date_add(date(my_date), INTERVAL -(EXTRACT(DAY FROM my_date))+1 day)) -1 end) -2 )/7)+1 end -- calculate week based on date
else FLOOR((EXTRACT(DAY FROM my_date)-1)/7)+1 -- for complete week
end
)
Is there a way to store a variable in bigQuery so my code can be less confusing to look? So far I've been reading bQ documentation and haven't found a way to store a variable inside function (for SQL UDF)
Any help would be appreciated, thanks!
Agree - it looks ugly!
So, I see you are reusing below expression eight(8) times
EXTRACT(DAYOFWEEK FROM date_add(date(my_date), INTERVAL -(EXTRACT(DAY FROM my_date))+1 day))
Consider below approach to refactor your initial code to use that expression only once into ABC alias/field and than reuse it in a "variable" fashion
create or replace function internal.week_in_month(my_date TIMESTAMP)
returns FLOAT64 as
((
select case
when (case when ABC = 1 then 7 else ABC -1 end) > 1 then -- check first day of month to decide if it's a complete week (starts on Monday)
case when EXTRACT(DAY FROM my_date) <= 7 then -- for incomplete week
case when
(case when EXTRACT(DAYOFWEEK FROM my_date) = 1 then 7 else EXTRACT(DAYOFWEEK FROM my_date)-1 end) - EXTRACT(DAY FROM my_date) =
(case when ABC = 1 then 7
else ABC -1 end) -1 then 1 -- incomplete week 1
else FLOOR(( EXTRACT(DAY FROM my_date) + (case when ABC = 1 then 7
else ABC -1 end) -2 )/7)+1 end -- calculate week based on date
else FLOOR(( EXTRACT(DAY FROM my_date) + (case when ABC = 1 then 7
else ABC -1 end) -2 )/7)+1 end -- calculate week based on date
else FLOOR((EXTRACT(DAY FROM my_date)-1)/7)+1 -- for complete week
end
from unnest([struct(EXTRACT(DAYOFWEEK FROM date_add(date(my_date), INTERVAL -(EXTRACT(DAY FROM my_date))+1 day)) as ABC)])
));
We can go further and apply same approach to EXTRACT(DAY FROM my_date) which is used six(6) times
create or replace function internal.week_in_month(my_date TIMESTAMP)
returns FLOAT64 as
((
select case
when (case when ABC = 1 then 7 else ABC -1 end) > 1 then -- check first day of month to decide if it's a complete week (starts on Monday)
case when XYZ <= 7 then -- for incomplete week
case when
(case when EXTRACT(DAYOFWEEK FROM my_date) = 1 then 7 else EXTRACT(DAYOFWEEK FROM my_date)-1 end) - XYZ =
(case when ABC = 1 then 7
else ABC -1 end) -1 then 1 -- incomplete week 1
else FLOOR(( XYZ + (case when ABC = 1 then 7
else ABC -1 end) -2 )/7)+1 end -- calculate week based on date
else FLOOR(( XYZ + (case when ABC = 1 then 7
else ABC -1 end) -2 )/7)+1 end -- calculate week based on date
else FLOOR((XYZ-1)/7)+1 -- for complete week
end
from unnest([struct(EXTRACT(DAY FROM my_date) as XYZ)]),
unnest([struct(EXTRACT(DAYOFWEEK FROM date_add(date(my_date), INTERVAL -(XYZ)+1 day)) as ABC)])
));
In above example - I was lazy enough and used just XYZ and ABC - in hope that you will use appropriate naming which will make our final code readable enough to be happy with
I was doing a MODE.com SQL practice question about date format.
The practice question is: Write a query that counts the number of companies acquired within 3 years, 5 years, and 10 years of being founded (in 3 separate columns). Include a column for total companies acquired as well. Group by category and limit to only rows with a founding date.
It uses two tables:
tutorial.crunchbase_companies_clean_date table, which includes information about all the companies, like company name, founded year, etc.
tutorial.crunchbase_acquisitions_clean_datetable, which includes the information about all the acquired companies, like acquired company name, acquired date, etc.
My code is:
SELECT companies.category_code,
COUNT(CASE WHEN acq.acquired_at_cleaned - companies.founded_at_clean:: timestamp <= '3 years' THEN 1 ELSE NULL END) AS less_than_3_years,
COUNT(CASE WHEN acq.acquired_at_cleaned - companies.founded_at_clean:: timestamp <= '5 years' THEN 1 ELSE NULL END) AS between_3_to_5_years,
COUNT(CASE WHEN acq.acquired_at_cleaned - companies.founded_at_clean:: timestamp <= '10 years' THEN 1 ELSE NULL END) AS within_10_years,
COUNT(1) AS total
FROM tutorial.crunchbase_companies_clean_date companies
JOIN tutorial.crunchbase_acquisitions_clean_date acq
ON companies.permalink = acq.company_permalink
WHERE companies.founded_at_clean IS NOT NULL
GROUP BY 1
ORDER BY total DESC
The result is:
My result
The answer query is:
SELECT companies.category_code,
COUNT(CASE WHEN acquisitions.acquired_at_cleaned <= companies.founded_at_clean::timestamp + INTERVAL '3 years'
THEN 1 ELSE NULL END) AS acquired_3_yrs,
COUNT(CASE WHEN acquisitions.acquired_at_cleaned <= companies.founded_at_clean::timestamp + INTERVAL '5 years'
THEN 1 ELSE NULL END) AS acquired_5_yrs,
COUNT(CASE WHEN acquisitions.acquired_at_cleaned <= companies.founded_at_clean::timestamp + INTERVAL '10 years'
THEN 1 ELSE NULL END) AS acquired_10_yrs,
COUNT(1) AS total
FROM tutorial.crunchbase_companies_clean_date companies
JOIN tutorial.crunchbase_acquisitions_clean_date acquisitions
ON acquisitions.company_permalink = companies.permalink
WHERE founded_at_clean IS NOT NULL
GROUP BY 1
ORDER BY 5 DESC
The result is:
The answer result
You can see in the screenshots that the results are very similar, but some numbers are different.
The only difference I can see between my query and the answer is in the COUNT statements, but I don't really see the difference, for example, between: acq.acquired_at_cleaned - companies.founded_at_clean:: timestamp <= '3 years' and acquisitions.acquired_at_cleaned <= companies.founded_at_clean::timestamp + INTERVAL '3 years'
I tried adding INTERVAL in my SELECT statement:
SELECT companies.category_code,
COUNT(CASE WHEN acq.acquired_at_cleaned - companies.founded_at_clean:: timestamp <= INTERVAL '3 years' THEN 1 ELSE NULL END) AS less_than_3_years,
COUNT(CASE WHEN acq.acquired_at_cleaned - companies.founded_at_clean:: timestamp <= INTERVAL '5 years' THEN 1 ELSE NULL END) AS between_3_to_5_years,
COUNT(CASE WHEN acq.acquired_at_cleaned - companies.founded_at_clean:: timestamp <= INTERVAL '10 years' THEN 1 ELSE NULL END) AS within_10_years,
COUNT(1) AS total
and remove the INTERVAL from the answer query:
SELECT companies.category_code,
COUNT(CASE WHEN acquisitions.acquired_at_cleaned <= companies.founded_at_clean::timestamp + '3 years'
THEN 1 ELSE NULL END) AS acquired_3_yrs,
COUNT(CASE WHEN acquisitions.acquired_at_cleaned <= companies.founded_at_clean::timestamp + '5 years'
THEN 1 ELSE NULL END) AS acquired_5_yrs,
COUNT(CASE WHEN acquisitions.acquired_at_cleaned <= companies.founded_at_clean::timestamp + '10 years'
THEN 1 ELSE NULL END) AS acquired_10_yrs,
COUNT(1) AS total
But the results are the same.
I tried to know the result of just the difference between the acquired_date and founded_date, to see if the value can be compared with INTERVAL. The result is in days, which looks promising to me.
The result
I try to give all the information for your consideration. Hope somebody could help. Thank you in advance!
My suggestion is to add/subtract the INTERVAL to/from one date/time and then compare with the other date/time. Don't subtract the date/times and then compare to a string literal. Your database seems to understand '3 YEARS' as 3 * 365 days, regardless of the actual number of days between someDateTime and someDateTime +/- '3 YEARS'. The actual number of days from year to year could be 365 or 366, depending on whether a leap year is crossed.
Here's a simple example of comparing with a specific interval, which also requires we know whether and how many leap years were crossed.
Fiddle
The test case:
WITH dates AS (
SELECT '2021-01-01'::date AS xdate
)
SELECT xdate - (xdate - INTERVAL '1' YEAR) AS diff
, xdate - (xdate - INTERVAL '1' YEAR) = '1 YEAR' AS b1
, xdate - (xdate - INTERVAL '1' YEAR) = '365 DAYS' AS b2
, xdate - (xdate - INTERVAL '1' YEAR) = '366 DAYS' AS b3
FROM dates
;
-- AND --
WITH dates AS (
SELECT '2021-01-01'::date AS xdate
)
SELECT xdate - (xdate - INTERVAL '1' YEAR) AS diff
, xdate - (xdate - INTERVAL '1' YEAR) = INTERVAL '1' YEAR AS b1
, xdate - (xdate - INTERVAL '1' YEAR) = INTERVAL '365 DAYS' AS b2
, xdate - (xdate - INTERVAL '1' YEAR) = INTERVAL '366 DAYS' AS b3
FROM dates
;
Result:
diff
b1
b2
b3
366 days
f
f
t
Fiddle
WITH dates AS (
SELECT '2021-01-01'::date AS xdate
)
, diff AS (
SELECT xdate - (xdate - INTERVAL '1' YEAR) AS diff
FROM dates
)
SELECT diff
, CASE WHEN diff = (366*24*60*60 * INTERVAL '1' SECOND)
THEN 1
END AS compare1
, 366*24*60*60 AS seconds
, CASE WHEN diff = (366*24*60*60 * INTERVAL '1' SECOND)
THEN 1
END AS compare2
, CASE WHEN diff = '31622400 SECONDS'
THEN 1
END AS compare3
FROM diff
;
The result:
diff
compare1
seconds
compare2
compare3
366 days
1
31622400
1
1
Original response:
The fiddle for PostgreSQL
The behavior shown here (below) is similar to the posted behavior.
The problem is the value generated isn't necessarily what you think.
Here's a test case in postgresql which might be representative of your issue.
Maybe this is related to leap year, where the number of days in a year isn't constant.
So it's probably safer to compare the dates rather than assume some number of days, which is probably the assumption <= '3 years' makes.
The test SQL:
WITH test (acquired_at_cleaned, founded_at_clean, n) AS (
SELECT current_date, current_date - INTERVAL '4' YEAR, 4 UNION
SELECT current_date, current_date - INTERVAL '3' YEAR, 3 UNION
SELECT current_date, current_date - INTERVAL '2' YEAR, 2 UNION
SELECT current_date, current_date - INTERVAL '1' YEAR, 1
)
, cases AS (
SELECT test.*
, CASE WHEN acquired_at_cleaned <= founded_at_clean::timestamp + INTERVAL '3' year
THEN 1 ELSE NULL
END AS acquired_3_yrs_case1
, CASE WHEN acquired_at_cleaned - founded_at_clean::timestamp <= '3 year'
THEN 1 ELSE NULL
END AS acquired_3_yrs_case2
, acquired_at_cleaned - founded_at_clean::timestamp AS x1
, acquired_at_cleaned - (n * INTERVAL '1' YEAR) AS x2
FROM test
)
SELECT acquired_at_cleaned AS acquired
, founded_at_clean AS founded
, n
, acquired_3_yrs_case1 AS case1
, acquired_3_yrs_case2 AS case2
, x1, x2
FROM cases
ORDER BY founded_at_clean
;
The result:
acquired
founded
n
case1
case2
x1
x2
2021-12-25
2017-12-25 00:00:00
4
null
null
1461 days
2017-12-26 00:00:00
2021-12-25
2018-12-25 00:00:00
3
1
null
1096 days
2018-12-26 00:00:00
2021-12-25
2019-12-25 00:00:00
2
1
1
731 days
2019-12-26 00:00:00
2021-12-25
2020-12-25 00:00:00
1
1
1
365 days
2020-12-26 00:00:00
Interesting result.
I have MS SQL function DATEDIFF
SELECT DATEDIFF(QQ, 0, '2018-09-05')
that returns 474(integer).
What is PostgreSQL equivalent of this function ?
MSSQL 0 is the date '1900-01-01' in DATEDIFF(QQ, 0, '2018-09-05'), that function will get the number of QUARTER from 1900-01-01 to 2018-09-05
But PostgreSQL does not have a QUARTER number function.
You can try to use
EXTRACT(QUARTER FROM TIMESTAMP '2018-09-05') to get this month QUARTER number.
date_part('year',age('2018-09-05','1900-01-01')) get year number between
'2018-09-05' and '1900-01-01'.
then do some calculation.
select (EXTRACT(QUARTER FROM TIMESTAMP '2018-09-05') +
date_part('year',age('2018-09-05','1900-01-01')) * 4) -1 QUARTER
Results:
| quarter |
|---------|
| 474 |
I think the current accepted is incorrect.
e.g. if you change date '1900-01-01' to '2017-07-01' you actually get 6 quarter diff.
But expected result should be:
include first and last quarter: 5
exclude first and last quarter: 3
exclude last quarter: 4
select
age('2018-09-05','2017-07-01') age,
(EXTRACT(QUARTER FROM TIMESTAMP '2018-09-05') +
date_part('year',age('2018-09-05','2017-07-01')) * 4) -1 QUARTER
Result:
age |quarter|
--------------------|-------|
1 year 2 mons 4 days| 6|
The reason is age('2018-09-05','2017-07-01') return 1 year 2 mons 4 days (maybe more than 12 month).
The answer i use to calculate number of quarter:
with params as (
select
'2017-07-01 00:00:00'::timestamp date_from,
'2018-09-05 00:00:00'::timestamp date_to
)
select
extract( quarter from date_from)::int date_from_quarter,
(extract( quarter from date_to))::int date_to_quarter,
age(date_to, date_from) date_diff,
(extract(year from date_to) - extract(year from date_from))::int how_many_yr,
-- exclude begin and end
(extract(year from date_to) - extract(year from date_from))::int* 4 - extract( quarter from date_from)::int + (extract( quarter from date_to))::int - 1 quarter_diff,
-- exclude begin quarter
(extract(year from date_to) - extract(year from date_from))::int* 4 - extract( quarter from date_from)::int + (extract( quarter from date_to))::int quarter_diff_include_current_quarter
from params
;
Result:
date_from_quarter|date_to_quarter|date_diff |how_many_yr|quarter_diff|quarter_diff_include_current_quarter|
-----------------|---------------|--------------------|-----------|------------|------------------------------------|
3| 3|1 year 2 mons 4 days| 1| 3| 4|
Its complicated on Postgres:
( (DATE_PART('year', '2018-09-05') - DATE_PART('year', '1900-01-01')) * 12
+ (DATE_PART('month', '2018-09-05') - DATE_PART('month', '1900-01-01'))
) / 4
Additional resources: topic
Let's say you have two dates in the example table: start_date and end_date
Here is how to get the difference between these dates in number of quarters:
select * (year(end_date)*4 + quarter(end_date)) - (year(start_date)*4 + quarter(start_date)) as quarter_diff
from example
I want to fetch last 12 months data from db, I have written a query for that but that only giving me count and month but not year means month related to which year.
My Sql :
Select count(B.id),date_part('month',revision_timestamp) from package AS
A INNER JOIN package_revision AS B ON A.revision_id=B.revision_id
WHERE revision_timestamp > (current_date - INTERVAL '12 months')
GROUP BY date_part('month',revision_timestamp)
it gives me output like this
month | count
-------+-------
7 | 21
8 | 4
9 | 10
but I want year with month like 7 - 2012, or year in other col, doesn't matter
I believe you wanted this:
SELECT to_char(revision_timestamp, 'YYYY-MM'),
count(b.id)
FROM package a
JOIN package_revision b ON a.revision_id = b.revision_id
WHERE revision_timestamp >
date_trunc('month', CURRENT_DATE) - INTERVAL '1 year'
GROUP BY 1
select
count(B.id),
date_part('year', revision_timestamp) as year,
date_part('month',revision_timestamp) as month
from package as A
inner join package_revision as B on A.revision_id=B.revision_id
where
revision_timestamp > (current_date - INTERVAL '12 months')
group by
date_part('year', revision_timestamp)
date_part('month', revision_timestamp)
or
select
count(B.id),
to_char(revision_timestamp, 'YYYY-MM') as month
from package as A
inner join package_revision as B on A.revision_id=B.revision_id
where
revision_timestamp > (current_date - INTERVAL '12 months')
group by
to_char(revision_timestamp, 'YYYY-MM')
Keep in mind that, if you filter by revision_timestamp > (current_date - INTERVAL '12 months'), you'll get range from current date in last year (so if today is '2013-09-04' you'll get range from '2012-09-04')