The problem with a table is that one option must apply to everyone. Parameter or other methods allow customized runs, so that your own special cases can still run...

Eirikur Eiriksson (9/15/2014)

---

ScottPletcher (9/15/2014)

---

All of those seem far too disruptive to me.

Two quick possibilities:

1) Add an optional parameter to the proc which requires a certain value or the proc just...

That GROUP BY is gonna kill your performance. See if the code below is close to what you need.

I've added clustered indexes on the temp tables prior to loading...

You can't consider missing index info in isolation. At the very least, you also need to look at index usage stats. Both of those are the minimum needed...

The big delay is likely in allocating and formatting log space, not in the identity number assignment, which is trivial.

Thus, before you run the big INSERT, pre-allocate sufficient log space...

All of those seem far too disruptive to me.

Two quick possibilities:

1) Add an optional parameter to the proc which requires a certain value or the proc just exits (or does...

Luis Cazares (9/11/2014)

---

Your little example is really nice, real life isn't like that.

Please find the errors in the following code:

INSERT INTO dbo.FactInternetSalesCopy

(ProductKey, OrderDateKey, DueDateKey, ShipDateKey, CustomerKey,...

thomashohner (9/9/2014)

---

I'm a noob but Predicates only evaluate to TRUE. Also NULL is not a value. Query 3 is correct as far as my very limited knowledge goes.

So it skips...

Most likely is that a database backup is in progress, and you can't alter the database files during a backup. Check running processes for a backup, dbcc check, or...

EXEC ('DELETE FROM onedb.stg.tab1') AT [linkserv]

You'll also need to decide if you're willing to allow the rounding to change the day/date or not. That is, if the time is, say, "20140908 23:59:59.670", do you...

The first, most obvious though is "don't do this". Why are you denormalizing one table's data into thousands of other tables? Why not just join to the original...

That's in case you get an entry like:

'Saturday'

Without the " + '0'", the entire string gets...

You could skip all non-numeric chars at the front of the string, viz:

SELECT

date_input,

SUBSTRING(date_input, PATINDEX('%[0-9]%', date_input + '0'), 30) AS adjusted_date_input

FROM (

...

select

g_a.id as GLAcct,

g_a.descr as GLDesc,

sum(case when g_l.transaction_date >= '20140101' AND g_l.transaction_date < '20140201' THEN g_l.amount_n ELSE 0 END) AS Jan,

--...

sum(case when g_l.transaction_date >= '20140701' AND g_l.transaction_date < '20140801' THEN g_l.amount_n...