Question

I have the following tables:

Table ROWS ( **RowId**, Title)

Table CELLS ( **CellId**, RowId, Title) 

Table ERRORS ( **ErrorId**, CellId, Title )

So rows can have cells and cells can have errors.

How do I write a query that retrieves the rows where the cells have no errors?

I know that something like this would work but I was wondering if it can be done in a more efficient manner :

QUERY1

SELECT ROWS.RowID FROM ROWS 
 JOIN CELLS ON ROWS.RowId = CELLS.RowID 
 JOIN ERRORS ON ERRORS.CellId = CELLS.CellId

QUERY2

SELECT * FROM ROWS 
WHERE ROWS.RowId NOT IN ( QUERY1 )** 

Any thoughts on how to do this in a more efficient manner ?

Answer 1

Some people have done research on this issue, see: http://explainextended.com/2009/09/18/not-in-vs-not-exists-vs-left-join-is-null-mysql/

The conclusion is: use LEFT JOIN / IS NULL or NOT IN. Don’t use NOT EXISTS; that is slower.

I think I would count the errors per row, and filter on that:

select cells.rowid from cells 
left join errors on errors.cellid=cells.cellid 
group by cells.rowid 
having count(errorid)=0;

This avoids all sub selects.