Regarding datatypes in SqlServer

Not too sure if you have read my reply. 

 

If you notice, you haven't really answered any of the questions asked by me. Not that you have to but it helps to decide the line of attack to resolve the issue.

 

Performance improvement is a huge topic and should be addressed from very many directions. Not all attempts will succeed. May be a combination of approaches need to applied.

 

Unfortunately, it is more of an art than science. So, we got to be patient too. In addition, at times we are bound by the design decisions made by someone and it is beyond our capacity to change them.

 

All we can do is strive to improve the situation as much as we can with least amount of damage and within our limits.

Hi,

 

We have big focus on performance and as it is now we must increase our performance in SQL queries. I talk about som queries that joins with a table wich is wrong created from the first time. En data is created with to column, one column for date (year,month,day) wich is a datetime datatype and another column where the time is stored (hour:minutes) wich is a char. When we do where clauses like (enddate+endtime) < @enddate sqlserver must do an implicit dataconversion and at that time sqlserver doesn't use our index. I try with explicit data conversion but didn't know if it is the solution.

Couple of ways to look at it. Because that's the way I look at it.

 

You didn't mention any timings of execution, so it is hard to comment on that. Have you already tried with and without where clause? What is the execution time difference? Is it truly painful as in 2 minutes, 3 minutes, etc. Share statistics, very important. Performance is a relative term.

 

Another thing is, to the best of my knowledge no indexes are used when calculations are part of where clause. I think it should be [column_name] = @parameter_value. Only then indexes are used.

 

Also what is volume? How many number of rows we are talking about here?

 

If you are desparate about performance, you may consider creating a materialized view on the underlying table with new datetime column calculated in the select itself. Depends on which version of SQL Server you are using and does it support materialized view.