A few minutes ago I visited this question on the main site:
How to design index for faster deletes
Link as it appears in my browser (IE8). I could not for the life of me see a comment option to ask a query so, for my sins, I posted it as an Answer. @gbn was kind enough to upvote it whilst I was...
If I count correctly, your function and question have at least four serious problems combined. 1. How do you try to call the function? 2. What error do you get? 3. you don't use the arguments in your function 4. the name is nonsense. — dezso18 secs ago
@James sigh we are all puzzled. The simple answer is because Walmart bought it, the not-so-simple answer is because their marketing team is full of idiots
I have to calculate shift hours of every staff member. I have written a query that gets the data from our OLTP system and returns the date, staff and their login logout time. Login and logout times are two separate transactions in the system but i want login and logout time in a single row. See b...
@ypercube well, no (the syntax itself is valid, but...)
SET check_function_bodies TO on;
CREATE OR REPLACE FUNCTION a() RETURNS boolean AS
$$
BEGIN
RETURN EXECUTE 'SELECT TRUE';
END;
$$
LANGUAGE plpgsql;
SELECT a();
and it is clear that the OP don't call the function in their question
@Marian another thing is that you wouldn't know when another user posts a duplicate answer, for instance. There could be a guy who opens the fiddle, copy the code and paste it as the answer. I would be prone to upvote that answer if I don't see the code on the other one.
@ypercube For the indexes, it would only be stores of PRIMARYs; at 1 million BIGINTs, it'd be only ~16mb. For rows, infrequent or unused ones would be removed from memory based upon open user sockets and time limits, with time limits on inactive connections. Only text will be buffered, definitely not images which require an extra user step to extract from the database. All site data will be stored in the database, will be written before updating c++ memory then users, and will be heavily normalized to allow for wiki-type editing, but the FOREIGNs will never be UPDATEd. — Gracchus7 mins ago
That's part of why I rarely answer. Most questions seem to fall into one of three categories: 1) Stuff other people will answer better than I would, 2) Incoherent stuff, 3) Sad people pasting mysqltuner output
@ypercube he's not doing cascade updates, the indexes being bigints will take up 16 bits apiece, so the index will be 16 bits * a million records or 16mb. He would like to use time-based caching so records are nearly constantly fresh. On first reading of a row (initial application code select) he doesn't want the binary fields from the database, those can be loaded individually later.
The last part about "and will be heavily normalized to allow for wiki-type editing" escapes me.
took me 5 seconds to figure what you were saying :)
The level of the migrated from SO questions today is awesome.
3
We're here for programming questions. Not server configuration. I recommend throwing your server off a very tall cliff. It should go much faster after the few seconds of freefall. — Marc B 1 hour ago
A few minutes ago I visited this question on the main site:
How to design index for faster deletes
Link as it appears in my browser (IE8). I could not for the life of me see a comment option to ask a query so, for my sins, I posted it as an Answer. @gbn was kind enough to upvote it whilst I was...
MSDN Article on removing duplicate rows with distinct
SQL Fiddle of the code
http://sqlfiddle.com/#!3/9a39a/1
