I have a Postgres materialized view that gets refreshed every night - a fairly large dataset. The first time someone tries to run it after a refresh, the query takes 9-10 minutes. After that, queries return in their usual 1-2 seconds.
SELECT * FROM MyMaterializedView
WHERE response_time BETWEEN '2022-01-01' AND '2022-02-01'
AND property_id = 123
...and there is an index on the Materialized View for the response_time and property_id columns.
My question is - is it normal for the initial running of a materialized view to take so much longer, and is there a way to handle this in the refresh to prevent the initial long-running query?
I can think of two possible reasons:
the data are not cached the first time you query the materialized view
the first query has to set the “hint bits” on all the rows it selects
You should set track_io_timing = on and run EXPLAIN (ANALYZE, BUFFERS) SELECT ... on your query to know for sure. If it is the first problem, you will see lots of buffers read and a high I/O time, in the second case you'll see lots of buffers dirtied or buffers written.
Related
So I have a query that I need a column with an aggregate within a determined frame, as example:
SUM(subquery1.value) OVER (
PARTITION BY subquery1.entity
ORDER BY subquery1.timestamp
ROWS BETWEEN 55 PRECEDING and CURRENT ROW
)
And so it works, gives me the expected result and it doesn't take up much time. Here's an EXPLAIN ANALYZE of it:
I know the statics are a bit off, but I'm happy with it.
The problem is that the 55 preceding rows is not really what I want, in fact what I actually want is closer to 50,000 and that's when it gets slower. If i bump the PRECEDING to 555 it already gets much slower, for higher values I haven't waited the run.
So I don't know whats happening, I have tried the same aggregate directly from a table using 300,000+ rows and it takes less than one second. This example is just over 50k rows so it should be much much faster.
If I have large amounts of data in a table defined like
CREATE TABLE sensor_values ( ts TIMESTAMPTZ(35, 6) NOT NULL,
value FLOAT8(17, 17) DEFAULT 'NaN' :: REAL NOT NULL,
sensor_id INT4(10) NOT NULL, );
Data comes in every minute for thousands of points. Quite often though I need to extract and work with daily values over years (On a web frontend). To aid this I would like a sensor_values_days table that only has the daily sums for each point and then I can use this for faster queries over longer timespans.
I don't want a trigger for every write to the db as I am afraid that would slow down the already bottle neck of writes to the db.
Is there a way to trigger only after so many rows have been inserted ?
Or perhaps an index and maintains a index of a sum of entries over days ? I don't think that is possible.
What would be the best way to do this. It would not have to be very up to date. Losing the last few hours or a day would not be an issue.
Thanks
What would be the best way to do this.
Install clickhouse and use AggregatingMergeTree table type.
With postgres:
Create per-period aggregate table. You can have several with different granularity, like hours, days, and months.
Have a cron or scheduled task run at the end of each period plus a few minutes. First, select the latest timestamp in the per-period table, so you know at which period to start. Then, aggregate all rows in the main table for periods that came after the last available one. This process will also work if the per-period table is empty, or if it missed the last update then it will catch up.
In order to do only inserts and no updates, you have to run it at the end of each period, to make sure it got all the data. You can also store the first and last timestamp of the rows that were aggregated, so later if you check the table you see it did use all the data from the period.
After aggregation, the "hour" table should be 60x smaller than the "minute" table, that should help!
Then, repeat the same process for the "day" and "month" table.
If you want up-to-date stats, you can UNION ALL the results of the "per day" table (for example) to the results of the live table, but only pull the current day out of the live table, since all the previous days's worth of data have been summarized into the "per day" table. Hopefully, the current day's data will be cached in RAM.
It would not have to be very up to date. Losing the last few hours or a day would not be an issue.
Also if you want to partition your huge table, make sure you do it before its size becomes unmanageable...
Materialized Views and a Cron every 5 minutes can help you:
https://wiki.postgresql.org/wiki/Incremental_View_Maintenance
In PG14, we will have INCREMENTAL MATERIALIZED VIEW, but for the moment is in devel.
i am struggling with my dashboard performance which runs queries on Redshift using JDBC driver.
the query is like -
select <ALIAS_TO_SCHEMA.TABLENAME>.<ANOTHER_COLUMN_NAME> as col_0_0_,
sum(<ALIAS_TO_SCHEMA.TABLENAME>.devicecount) as col_1_0_ from <table_schema>.<table_name> <ALIAS_TO_SCHEMA.TABLENAME> where <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$1
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$2
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$3
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$4
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$5
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$6
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$7
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$8
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$9
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$10
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$11
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$12
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$13
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$14
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$15
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$16
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$17
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$18
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$19
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$20
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$21
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$22
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$23
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$24
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$25
or <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME>=$26
or ........
The For dashboard we use Spring, Hibernate ( I am not 100% sure about it though ).
But the query might sometimes stretch till $1000 + according to the filters/options being selected on the UI.
But the problem we are seeing is - The First Time this query is being run by the reports, it takes more than 40 sec - 60 seconds for the response. After the first time , the query runs quite fast and takes only few seconds to run.
We initially suspected there must be something wrong with redshift caching , but it turns out that , Even simple queries like these ( But Huge ) takes considerable time to COMPILE, which is clear when we look into the svl_compile table which shows this query was compiled in over 35 seconds.
What should I do to handle such issues ?
Recommend restructuring the query generated by your dashboard to use an IN list. Redshift should be able to reuse the already compiled query segments for different length IN lists.
Note that IN lists with less than 10 values will still be evaluated as OR. https://docs.aws.amazon.com/redshift/latest/dg/r_in_condition.html#r_in_condition-optimization-for-large-in-lists
SELECT <ALIAS_TO_SCHEMA.TABLENAME>.<ANOTHER_COLUMN_NAME> as col_0_0_
, SUM(<ALIAS_TO_SCHEMA.TABLENAME>.devicecount) AS col_1_0_
FROM <table_schema>.<table_name> <ALIAS_TO_SCHEMA.TABLENAME>
WHERE <ALIAS_TO_SCHEMA.TABLENAME>.<COLUMN_NAME> IN ( $1,$2,$3,$4,$5,$6,$7,$8,$9,$10,$11 … $1000 )
;
I'm looking at a postgres system with tables containing 10 or 100's of millions of rows, and being fed at a rate of a few rows per second.
I need to do some processing on the rows of these tables, so I plan to run some simple select queries: select * with a where clause based on a range (each row contains a timestamp, that's what I'll work with for ranges). It may be a "closed range", with a start and an end I know are contained in the table, and I know no new data will fall into the range, or an open range : ie one of the range boundary might not be "in the table yet" and rows being fed in the table might thus fall in that range.
Since the response will itself contains millions of rows, and the processing per row can take some time (10s of ms) I'm fully aware I'll use a cursor and fetch, say, a few 1000 rows at a time. My question is:
If I run an "open range" query: will I only get the result as it was when I started the query, or will new rows being inserted in the table that fall in the range while I run my fetch show up ?
(I tend to think that no I won't see new rows, but I'd like a confirmation...)
updated
It should not happen under any isolation level:
https://www.postgresql.org/docs/current/static/transaction-iso.html
but Postgres insures it only in Serializable isolation
Well, I think when you make a query, that means you create a new transaction and it will not receive/update data from any other transaction until it commit.
So, basically "you only get the result as it was when you started the query"
When I execute a query and right click in the results area, I get a pop-up menu with the following options:
Save Grid as Report ...
Single Record View ...
Count Rows ...
Find/Highlight ...
Export ...
If I select "Count Rows", is there a way to interrupt the operation if it starts taking too long?
No, you don't seem to be able to.
When you select Count Rows from the context menu, it runs the count on the main UI thread, hanging the whole UI, potentially for minutes or hours.
It's best not to use that feature - just put select count (*) from ( < your query here>) which it runs properly on separate thread which can be cancelled.
You can open an new instance of sql developer and kill the session counting the rows.
I do suggest using the select count(*) query though as it is less painful in the long run.