I have two tables each with 100 million+ rows, table_1 and tabke_2.
We insert 60,000+ rows with the same date of "today" in one column of each table. This "date" field is indexed in each of the two tables.
We're doing this insert every day.
Following the insert, if we run a query
select count(*)
from ((select field1 from table_1 where date_field = 'yyyy-mm-dd' --**yesterday's date** ) a
INNER JOIN
(select field1 from table_2 where date_field = 'yyyy-mm-dd' --**yesterday's date** ) b
ON a.field1 = b.field1) c
runs in 1 second
select count(*)
from ((select field1 from table_1 where date_field = 'yyyy-mm-dd' --**today's date** ) a
INNER JOIN
(select field1 from table_2 where date_field = 'yyyy-mm-dd' --**today's date** ) b
ON a.field1 = b.field1) c
runs in 6 hours!
Tomorrow, this query will run in 1 second, and the next day's date query runs for 6 hours.
I'm totally puzzled. Why does the same query runs for 1 second, but the recently inserted data query runs for 6 hours? The next day the 6 hours query runs in 1 second, and that day's date query runs for 6 hours...
I would check the explain plans for the queries, and I suspect that statistics are being automatically gathered after your slow-performing query that change the execution plan by the next day.
Edit: Oh so the fix would be to invoke analyse after loading new data.
Related
This statement works in SQLite, but not in Postgres:
SELECT A.*, B.*
FROM Readings A
LEFT JOIN Offsets B ON A.MeterNum = B.MeterNo AND A.DateTime > B.TimeDate
WHERE A.MeterNum = 1
GROUP BY A.DateTime
ORDER BY A.DateTime DESC
The Readings table contains electric submeter readings each with a date stamp. The Offsets table holds an adjustment that the user enters after a failed meter is replaced with a new one that starts again at zero. Without the Group By statement the query returns a line for each meter reading with each prior adjustment made before the reading date while I only want the last adjustment.
All the docs I've seen on Group By for Postgres indicate I should be including an aggregate function which I don't need and can't use (The Reading column contains the Modbus string returned from the meter).
Just pick the latest reading in a derived table. In Postgres this can be done quite efficiently using distinct on ()
SELECT A.*, B.*
FROM readings A
left join (
select distinct on (meterno) o.*
from offsets o
order by o.meterno, o.timedate desc
) B ON A.MeterNum = B.MeterNo AND A.DateTime > B.TimeDate
WHERE A.meternum = 1
ORDER BY A.DateTime DESC
distinct on () will only return one row per meterno and this is the "latest" row due to the order by ... , timedate desc
The query might even be faster by pushing the condition on datetime > timedate into the derived table using a lateral join:
SELECT A.*, B.*
FROM readings A
left join lateral (
select distinct on (meterno) o.*
from offsets o
where a.datetime > o.timedeate
order by o.meterno, o.timedate desc
) B ON A.MeterNum = B.MeterNo
WHERE A.meternum = 1
ORDER BY A.DateTime DESC
I have a table with duplicate entries and the objective is to get the distinct entries based on the latest time stamp.
In my case 'serial_no' will have duplicate entries but I select unique entries based on the latest time stamp.
Below query is giving me the unique results with the latest time stamp.
But my concern is I need to get the total of unique entries.
For example assume my table has 40 entries overall. With the below query I am able to get 20 unique rows based on the serial number.
But the 'total' is returned as 40 instead of 20.
Any help on this pls?
SELECT
*
FROM
(
SELECT
DISTINCT ON (serial_no) id,
serial_no,
name,
timestamp,
COUNT(*) OVER() as total
FROM
product_info
INNER JOIN my.account ON id = accountid
WHERE
lower(name) = 'hello'
ORDER BY
serial_no,
timestamp DESC OFFSET 0
LIMIT
10
) AS my_info
ORDER BY
serial_no asc
product_info table intially has this data
serial_no name timestamp
11212 pulp12 2018-06-01 20:00:01
11213 mango 2018-06-01 17:00:01
11214 grapes 2018-06-02 04:00:01
11215 orange 2018-06-02 07:05:30
11212 pulp12 2018-06-03 14:00:01
11213 mango 2018-06-03 13:00:00
After the distict query I got all unique results based on the latest
timestamp:
serial_no name timestamp total
11212 pulp12 2018-06-03 14:00:01 6
11213 mango 2018-06-03 13:00:00 6
11214 grapes 2018-06-02 04:00:01 6
11215 orange 2018-06-02 07:05:30 6
But total is appearing as 6 . I wanted the total to be 4 since it has
only 4 unique entries.
I am not sure how to modify my existing query to get this desired
result.
Postgres supports COUNT(DISTINCT column_name), so if I have understood your request, using that instead of COUNT(*) will work, and you can drop the OVER.
What you could do is move the window function to a higher level select statement. This is because window function is evaluated before distinct on and limit clauses are applied. Also, you can not include DISTINCT keyword within window functions - it has not been implemented yet (as of Postgres 9.6).
SELECT
*,
COUNT(*) OVER() as total -- here
FROM
(
SELECT
DISTINCT ON (serial_no) id,
serial_no,
name,
timestamp
FROM
product_info
INNER JOIN my.account ON id = accountid
WHERE
lower(name) = 'hello'
ORDER BY
serial_no,
timestamp DESC
LIMIT
10
) AS my_info
Additionally, offset is not required there and one more sorting is also superfluous. I've removed these.
Another way would be to include a computed column in the select clause but this would not be as fast as it would require one more scan of the table. This is obviously assuming that your total is strictly connected to your resultset and not what's beyond that being stored in the table, but gets filtered out.
select count(*), serial_no from product_info group by serial_no
will give you the number of duplicates for each serial number
The most mindless way of incorporating that information would be to join in a sub query
SELECT
*
FROM
(
SELECT
DISTINCT ON (serial_no) id,
serial_no,
name,
timestamp,
COUNT(*) OVER() as total
FROM
product_info
INNER JOIN my.account ON id = accountid
WHERE
lower(name) = 'hello'
ORDER BY
serial_no,
timestamp DESC OFFSET 0
LIMIT
10
) AS my_info
join (select count(*) as counts, serial_no from product_info group by serial_no) as X
on X.serial_no = my_info.serial_no
ORDER BY
serial_no asc
I am trying to figure out the HQL equivalent of my query that has 2 subqueries, what I'm trying to do is I am getting the max amount for the past 6 months and then group them by month, and then I will get the average of the past 6 month result. And since the rows have version column, I also need to get maxed version for that specific row.
Here is my query, I'm using postgres by the way. Any help would be appreciated as I'm really having a hard time. Thanks in advance.
select avg(amount1) as maxField1 from
(
select max(amount1) as amount1 from table1 a
where a.id = :id
and a.date between :startDate
and :endDate
and a.version =
(
select max(b.version) from table1 b
where a.id = b.id
and a.date = b.date
)
group by to_char(a.date, 'YYYYMM')
);
I have a very simpl postgres (9.3) query that looks like this:
SELECT a.date, b.status
FROM sis.table_a a
JOIN sis.table_b b ON a.thing_id = b.thing_id
WHERE EXTRACT(MONTH FROM a.date) = 06
AND EXTRACT(YEAR FROM a.date) = 2015
Some days of the month of June do not exist in table_a and thus are obviously not joined to table_b. What is the best way to create records for these not represented days and assign a placeholder (e.g. 'EMPTY') to their 'status' column? Is this even possible to do using pure SQL?
Basically, you need LEFT JOIN and it looks like you also need generate_series() to provide the full set of days:
SELECT d.date
, a.date IS NOT NULL AS a_exists
, COALESCE(b.status, 'status_missing') AS status
FROM (
SELECT date::date
FROM generate_series('2015-06-01'::date
, '2015-06-30'::date
, interval '1 day') date
) d
LEFT JOIN sis.table_a a USING (date)
LEFT JOIN sis.table_b b USING (thing_id)
ORDER BY 1;
Use sargable WHERE conditions. What you had cannot use a plain index on date and has to default to a much more expensive sequential scan. (There are no more WHERE conditions in my final query.)
Aside: don't use the basic type name (and reserved word in standard SQL) date as identifier.
Related (2nd chapter):
PostgreSQL: running count of rows for a query 'by minute'
So far I have come up with the below:
WHERE (extract(month FROM orders)) =
(SELECT min(extract(month from orderdate))
FROM orders)
However, that will consequently return zero to many rows, and in my case, many, because many orders exist within that same earliest (minimum) month, i.e. 4th February, 9th February, 15th Feb, ...
I know that a WHERE clause can contain multiple columns, so why wouldn't the below work?
WHERE (extract(day FROM orderdate)), (extract(month FROM orderdate)) =
(SELECT min(extract(day from orderdate)), min(extract(month FROM orderdate))
FROM orders)
I simply get: SQL Error: ORA-00920: invalid relational operator
Any help would be great, thank you!
Sample data:
02-Feb-2012
14-Feb-2012
22-Dec-2012
09-Feb-2013
18-Jul-2013
01-Jan-2014
Output:
02-Feb-2012
14-Feb-2012
Desired output:
02-Feb-2012
I recreated your table and found out you just messed up the brackets a bit. The following works for me:
where
(extract(day from OrderDate),extract(month from OrderDate))
=
(select
min(extract(day from OrderDate)),
min(extract(month from OrderDate))
from orders
)
Use something like this:
with cte1 as (
select
extract(month from OrderDate) date_month,
extract(day from OrderDate) date_day,
OrderNo
from tablename
), cte2 as (
select min(date_month) min_date_month, min(date_day) min_date_day
from cte1
)
select cte1.*
from cte1
where (date_month, date_day) = (select min_date_month, min_date_day from cte2)
A common table expression enables you to restructure your data and then use this data to do your select. The first cte-block (cte1) selects the month and the day for each of your table rows. Cte2 then selects min(month) and min(date). The last select then combines both ctes to select all rows from cte1 that have the desired month and day.
There is probably a shorter solution to that, however I like common table expressions as they are almost all the time better to understand than the "optimal, shortest" query.
If that is really what you want, as bizarre as it seems, then as a different approach you could forget the extracts and the subquery against the table to get the minimums, and use an analytic approach instead:
select orderdate
from (
select o.*,
row_number() over (order by to_char(orderdate, 'MMDD')) as rn
from orders o
)
where rn = 1;
ORDERDATE
---------
01-JAN-14
The row_number() effectively adds a pseudo-column to every row in your original table, based on the month and day in the order date. The rn values are unique, so there will be one row marked as 1, which will be from the earliest day in the earliest month. If you have multiple orders with the same day/month, say 01-Jan-2013 and 01-Jan-2014, then you'll still only get exactly one with rn = 1, but which is picked is indeterminate. You'd need to add further order by conditions to make it deterministic, but I have no idea what you might want.
That is done in the inner query; the outer query then filters so that only the records marked with rn = 1 is returned; so you get exactly one row back from the overall query.
This also avoids the situation where the earliest day number is not in the earliest month number - say if you only had 01-Jan-2014 and 02-Feb-2014; comparing the day and month separately would look for 01-Feb-2014, which doesn't exist.
SQL Fiddle (with Thomas Tschernich's anwer thrown in too, giving the same result for this data).
To join the result against your invoice table, you don't need to join to the orders table again - especially not with a cross join, which is skewing your results. You can do the join (at least) two ways:
SELECT
o.orderno,
to_char(o.orderdate, 'DD-MM-YYYY'),
i.invno
FROM
(
SELECT o.*,
row_number() over (order by to_char(orderdate, 'MMDD')) as rn
FROM orders o
) o, invoices i
WHERE i.invno = o.invno
AND rn = 1;
Or:
SELECT
o.orderno,
to_char(o.orderdate, 'DD-MM-YYYY'),
i.invno
FROM
(
SELECT orderno, orderdate, invno
FROM
(
SELECT o.*,
row_number() over (order by to_char(orderdate, 'MMDD')) as rn
FROM orders o
)
WHERE rn = 1
) o, invoices i
WHERE i.invno = o.invno;
The first looks like it does more work but the execution plans are the same.
SQL Fiddle with your pastebin-supplied query that gets two rows back, and these two that get one.