Oracle SQL Month Statement Generation - sql

Question

I am having performance issue on a set of SQLs to generate current month's statement in realtime.
Customers will purchase some goods using points from an online system, and the statement containing "open_balance", "point_earned", "point_used", "current_balance" should be generated.
The following shows the shortened schema :
//~200k records
customer: {account_id:string, create_date:timestamp, bill_day:int} //totally 14 fields
//~250k records per month, kept for 6 month
history_point: {point_id:long, account_id:string, point_date:timestamp, point:int} //totally 9 fields
//each customer have maximum of 12 past statements kept
history_statement: {account_id:string, open_date:date, close_date:date, open_balance:int, point_earned:int, point_used:int, close_balance:int} //totally 9 fields
On every bill day, the view should automatically create a new month statement.
i.e. If bill_day is 15, then transaction done on or after 16 Dec 2013 00:00:00 should belongs to new bill cycle of 16 Dec 2013 00:00:00 - 15 Jan 2014 23:59:59
I tried the approach described below,
Calculate the last close day for each account (in materialized view, so that it update only after there is new customer or past month statement inserted into history_statement)
Generate a record for each customer each month that I need to calculate (Also in materialized view)
Sieve the point record for only point records within the date that I will calculate (This takes ~0.1s only)
Join 2 with 3 to obtain point earned and used for each customer each month
Join 4 with 4 on date less than open date to sum for open and close balance
6a. Select from 5 where open date is less than 1 month old as current balance (these are not closed yet, and the point reflect the point each customer own now)
6b. All the statements are obtained by union of history_statement and 5
On a development server, the average response time (200K customer, 1.5M transactions in current month) is ~3s which is pretty slow for web application, and on the testing server, where resources are likely to be shared, the average response time (200K customer, ~200k transaction each month for 8 months) is 10-15s.
Does anyone have some idea on writing a query with better approach or to speed up the query?
Related SQL:
2: IV_STCLOSE_2_1_T(Materialized view)
3: IV_STCLOSE_2_2_T (~0.15s)
SELECT ACCOUNT_ID, POINT_DATE, POINT
FROM history_point
WHERE point_date >= (
SELECT MIN(open_date)
FROM IV_STCLOSE_2_1_t
)
4: IV_STCLOSE_3_T (~1.5s)
SELECT p0.account_id, p0.open_date, p0.close_date, COALESCE(SUM(DECODE(SIGN(p.point),-1,p.point)),0) AS point_used, COALESCE(SUM(DECODE(SIGN(p.point),1,p.point)),0) AS point_earned
FROM iv_stclose_2_1_t p0
LEFT JOIN iv_stclose_2_2_t p
ON p.account_id = p0.account_id
AND p.point_date >= p0.open_date
AND p.point_date < p0.close_date + INTERVAL '1' DAY
GROUP BY p0.account_id, p0.open_date, p0.close_date
5: IV_STCLOSE_4_T (~3s)
WITH t AS (SELECT * FROM IV_STCLOSE_3_T)
SELECT t1.account_id AS STAT_ACCOUNT_ID, t1.open_date, t1.close_date, t1.open_balance, t1.point_earned AS point_earn, t1.point_used , t1.open_balance + t1.point_earned + t1.point_used AS close_balance
FROM (
SELECT v1.account_id, v1.open_date, v1.close_date, v1.point_earned, v1.point_used, COALESCE(sum(v2.point_used + v2.point_earned),0) AS OPEN_BALANCE
FROM t v1
LEFT JOIN t v2
ON v1.account_id = v2.account_id
AND v1.OPEN_DATE > v2.OPEN_DATE
GROUP BY v1.account_id, v1.open_date, v1.close_date, v1.point_earned, v1.point_used
) t1

Answer 1

It turns out to be that in IV_STCLOSE_4_T
WITH t AS (SELECT * FROM IV_STCLOSE_3_T)
is problematic.
At first thought WITH t AS would be faster as IV_STCLOSE_3_T is only evaluated once, but it apparently forced materializing the whole IV_STCLOSE_3_T, generating over 200k records despite I only need at most 12 of them from a single customer at any time.
With the above statement removed and appropriately indexing account_id, the cost reduced from over 500k to less than 500.