SQL IN查询性能-更好地拆分或不拆分

Question

I get up to 1000 id's from another server to display them for visitors so I have to use the IN query like:

SELECT * FROM `table` WHERE `id` IN (23221, 42422, 2342342....) // and so on, up to 1000

Let's say 1/3 of the visitors will watch though all of the 1000 id's while 2/3 of the them will only watch the first 50.

What would be a better for performance/workload, one query for all the 1000 id's or split them into like 20 queries so 50 id's each? So when the first 50 have been watched, query for the next 50 etc.

EDIT:

I don't need to use LIMIT when spliting, which means the id's in the query would be 50 max. So what's better, one query with 1000 id's at once or 20 queries each 50 id's?

EDIT:

Ok I ask it shortly and more directly: are 1000 id's in one query not too much? I have read here How to optimize an SQL query with many thousands of WHERE clauses that tons of WHERE/OR are bad??

Answer 1

Let's say 1/3 of the visitors will watch though all of the 1000 id's while 2/3 of the them will only watch the first 50.

Since you want to optimize your response as you assumed how visitors will treat it.

What would be a better for performance/workload, one query for all the 1000 id's or split them into like 20 queries so 50 id's each? So when the first 50 have been watched, query for the next 50 etc.

Yes, you are correct you should limit the return response. This is one example of how you can implement your requirement (I don't know much mysql but this is how you could get desired result).

SELECT * FROM `table` WHERE `id` IN (23221, 42422, 2342342....)
order by `id`
LIMIT 10 OFFSET 10

if it was SQL SERVER :

  create stored proc sp_SomeName
  @id varchar(8000)
  @skip int,
  @take int
  as
  begin
     SELECT * FROM some_table WHERE id IN (23221, 42422, 2342342....)
     order by id
     OFFSET @skip ROWS --if 0 then start selecting from 0 
     FETCH NEXT @take ROWS ONLY --if 10 then this is the max returning limit
  end  

what above query will do is : It will get all the data of the posted ids, then it will order by id in ascending order. Then from their it will choose just first 10/50/100, next time, it will choose the next 10/50/100 or whatever your take choice is and skip choice is. Hope this helps man :)

Answer 2

You can look at the answer provided here: MySQL Data - Best way to implement paging?

With the LIMIT statement you can return only a portion of the result. And by changing the parameters in the LIMIT statement, you can re-use the query.

Do know that unless you use an 'ORDER BY', an SQL server does not always return the same records. In other words, should a record by unavailable to read due to an update that occurs, while the database-server can read the next record, it will fetch the next record (to give a result as soon as possible). I do not know for sure if the LIMIT forces a database-server to take some sort of order into consideration (I am not that familiar with MySql).