2017-09-14 15:09 GMT+02:00 Pavel Stehule <pavel.steh...@gmail.com>:
2017-09-14 14:59 GMT+02:00 Frank Millman <fr...@chagford.com>:
Pavel Stehule wrote:
2017-09-14 10:14 GMT+02:00 Frank Millman <fr...@chagford.com>:
Hi all
This is a follow-up to a recent question I posted regarding a slow query.
I thought that the slowness was caused by the number of JOINs in the query, but
with your assistance I have found the true reason. I said in the previous
thread that the question had become academic, but now that I understand things
better, it is no longer academic as it casts doubt on my whole approach.
I have split my AR transaction table into three physical tables –
ar_tran_inv, ar_tran_crn, ar_tran_rec. I will probably add others at some
point, such as ar_tran_jnl.
I then create a VIEW to view all transactions combined. The view is
created like this -
CREATE VIEW ar_trans AS
SELECT ‘ar_inv’ AS tran_type, row_id AS tran_row_id, tran_number ...
FROM ar_tran_inv WHERE posted = ‘1’
UNION ALL
SELECT ‘ar_crn’ AS tran_type, row_id AS tran_row_id, tran_number ...
FROM ar_tran_crn WHERE posted = ‘1’
UNION ALL
SELECT ‘ar_rec’ AS tran_type, row_id AS tran_row_id, tran_number ...
FROM ar_tran_rec WHERE posted = ‘1’
I have another table called ‘ar_trans_due’, to keep track of outstanding
transactions. All of the three transaction types generate entries into this
table. To identify the source of the transaction, I have created columns in
ar_trans_due called ‘tran_type’ and ‘tran_row_id’. After inserting a row into
‘ar_tran_inv’, I invoke this -
INSERT INTO ar_trans_due (tran_type, tran_row_id, ...) VALUES
(‘ar_inv’, ar_tran_inv.row_id, ...), and similar for the other transaction
types. It is handled by a Python program, and it all happens within a
transaction.
When I view a row in ar_trans_due, I want to retrieve data from the
source transaction, so I have this -
SELECT * FROM ar_trans_due a
LEFT JOIN ar_trans b ON b.tran_type = a.tran_type AND b.tran_row_id =
a.tran_row_id
I understand that PostgreSQL must somehow follow a path from the view
‘ar_trans’ to the physical table ‘ar_tran_inv’, but I assumed it would execute
the equivalent of SELECT * FROM ar_tran_inv WHERE row_id = a.tran_row_id AND
posted = ‘1’.
If this was the case, it would be an indexed read, and very fast.
Instead, according to EXPLAIN, it performs a sequential scan of the
‘ar_tran_inv’ table.
It also scans ‘ar_tran_crn’ and ‘ar_tran_rec’, but EXPLAIN shows that it
uses a Bitmap Heap Scan on those. I assume that is because the tables are
currently empty.
Is this analysis correct?
please, send EXPLAIN ANALYZE result :)
I tried to reduce this to its simplest form.
Here is a SQL statement -
SELECT *
FROM ccc.ar_trans_due a
LEFT JOIN ccc.ar_trans b ON
b.tran_type = a.tran_type AND b.tran_row_id = a.tran_row_id
WHERE a.row_id = 1
ar_trans_due is a physical table, ar_trans is a view.
It takes about 28ms. Here is the explain - https://explain.depesz.com/s/8YY
> The PostgreSQL cannot to push join - in slow case, the UNIONS should be done
> first - and it requires full scan ar_tran_inv - used filter (posted AND
> (deleted_id = 0) is not too effective - maybe some composite or partial index
> helps.
>
> The fast query doesn't contains unions - so there are bigger space for
> optimizer - ar_tran_inv is filtered effective - by primary key.
>
> So main problem is impossible to push information a.row_id = 1 to deep to
> query.
>
Sorry for banging on about this, but someone might be interested in the
following timings.
The only solution I could find was to ‘denormalise’ (if that is a word) and
create additional columns on ar_trans_due for cust_row_id and tran_date, to
avoid using any joins.
Once I had done that, I could run my query two ways –
1. using the newly created columns
2. as before, using a join to the view, which in turn retrieved data from the
underlying tables.
This was a more complex query than the example above – details available on
request.
Here are the timings for running the query on identical data sets using
Postgresql, Sql Server, and Sqlite3 -
PostgreSQL -
Method 1 - 0.28 sec
Method 2 – 1607 sec, or 26 minutes
Sql Server -
Method 1 – 0.33 sec
Method 2 – 1.8 sec
Sqlite3 -
Method 1 – 0.15 sec
Method 2 – 1.0 sec
It seems that Sql Server and Sqlite3 are able to analyse the ‘join’, and
execute an indexed read against the underlying physical tables.
Frank
