MARIADB variables in SELECT statements POSTGRES equivelent - postgresql

Question

In mariadb/mysql I can use variables in the following way to perform calculations in a select statement. In this simple instance, I create a range of dates and use variables to calculate a simple opening balance and closing balance, with a payment of 10 each day.
with RECURSIVE dates as (
select '2017-11-01' as `dt`
union all
select adddate(dt, INTERVAL 1 DAY)
from dates
where dt < CURDATE()
)
select
#vardate:=d.dt
, #openbal
, #payment:= 10
, #closebal:= #openbal+#payment
, #openbal:=#closebal
from dates d;
gives the results....
╔══════════════╦════════════╦═══════════════╦═════════════╗
║ "ac_date" ║ "open_bal" ║ "trans_total" ║ "close_bal" ║
╠══════════════╬════════════╬═══════════════╬═════════════╣
║ "2017-11-01" ║ "0" ║ "10" ║ "10" ║
║ "2017-11-02" ║ "10" ║ "10" ║ "20" ║
║ "2017-11-03" ║ "20" ║ "10" ║ "30" ║
...
Using this technique I can perform simple calculations on the fly in a select statement. My question is, is it possible to use variables in this way in a PL/pgSQL function or is there an alternative method I am overlooking?

Answer 1

I am not entirely sure how that statement works in MariaDB, but this seems to do the same thing:
with vars (openbal, payment) as (
values (0, 10)
), balance as (
select t.dt::date as ac_date,
openbal,
payment,
sum(payment) over (order by t.dt) as close_bal
from vars,
generate_series(date '2017-11-01', current_date, interval '1' day ) as t(dt)
)
select ac_date,
openbal + lag(close_bal) over (order by ac_date) as open_bal,
payment,
close_bal
from balance;
In general to get a running total you use sum() over (order by ...) in SQL. To access values from a previous row, you use the lag() function.
The two CTEs are needed because window functions can't be nested.
To generate a list of rows, use generate_series() in Postgrse.