is it cache or something else

Question

is it cache or something else

May 13, 2025 at 10:18 am

This was removed by the editor as SPAM

Viewing 7 posts - 16 through 21 (of 21 total)

You must be logged in to reply to this topic. Login to reply

Lisa Slater Nicholls Hall of Fame Points: 3461 More actions · Answer 1

Hi again Stan, sorry I didn't help. The only other thing I can think of -- and it may make no difference -- is to remove the two fields from the RDL dataset, using DataSet properties window -> Fields panel , so that even though the 2 columns remain in the view the DataSet isn't evaluating them. You mentioned they're not needed for the report, right?

Also, I'm curious to know whether the RDL has them properly typed. You'll need to open the RDL in a text editor or XML editor to see this. I think you also mentioned that the two fields were added to the view later. If this is a case where the dataset query is basically select * from the_view, maybe this later addition has them added wrong. (System.Int.. versus System.Decimal)

Come to think of it, if this is a select * from the_view and you don't need all the fields, I'd change the RDL query to explicitly choose the set of fields you do need. Be careful of casing, of course -- you will need to exactly match the casing of the columns you do need from what is currently "seen" in the fields list via select * -- but if you do this it's both more efficient (why drag more data than you need?) and may solve your problem. I would totally try this first.

This reply was modified 1 week, 1 day ago by Lisa Slater Nicholls.

Jo Pattyn SSC-Dedicated Points: 32415 More actions · Answer 2

Sometimes SQL Servers filters very late

and it is possible to have divide by zero on rows that aren't relevant

because they are done before SQL Server decides to filter them out.

SELECT B

FROM A

WHERE A.COL = 'Z'

SELECT *
FROM
(
SELECT
B
,CrazyCalc
,COL
)
WHERE COL='Z'

Jo Pattyn SSC-Dedicated Points: 32415 More actions · Answer 3

Sometimes SQL Servers filters very late

and it is possible to have divide by zero on rows that aren't relevant

because they are done before SQL Server decides to filter them out.

SELECT B

FROM A

WHERE A.COL = 'Z'

SELECT *
FROM
(
SELECT
B
,CrazyCalc
,COL
FROM A
) ABC
WHERE COL='Z'

stan Ten Centuries Points: 1105 More actions · Answer 4

thx lisa, yep that was the first thing i did. I set up a separate , temp / workaround copy of the reports without a reference to the offending fields. and told the users to use those for the time being. its not a select *.

stan Ten Centuries Points: 1105 More actions · Answer 5

thx jo, that is good to know and from what ive seen , the most plausible explanation so far. i dont know that i'll spend the money to know for sure but will pass it along to some folks who have offered to be a second pair of eyes on this.

Lisa Slater Nicholls Hall of Fame Points: 3461 More actions · Answer 6

I set up a separate , temp / workaround copy of the reports without a reference to the offending fields.

I guess I am not understanding why this is a "temporary" fix, if the report doesn't need those columns?

I do agree that, assuming the offending view is used by more things than this report, the view itself should be fixed... and if this means going "deeper" into the views feeding the final view, then *they* should be fixed, but obviously nobody here can help you with that without seeing more of the code.

Good luck, and I hope the other "eyes on" in your company can do that work with you successfully, unless you can share the full code here. Sometimes, just in the act of providing code to reproduce the problem (using system tables as standins for the originals, but preserving the logic as you go), you will find the issue yourself. 🙂