762 reputation
1719
bio website
location
age
visits member for 3 years, 6 months
seen yesterday

Jun
22
comment How to gracefully handle a source database restore from backup in a data warehouse
@MarkStorey-Smith: I think a rebuild is a possibility; is that how this is typically handled in the field?
Dec
14
comment ETL: extracting from 200 tables - SSIS data flow or custom T-SQL?
BIML looks very interesting. I was also considering creating each data flow as a separate package and then having them called through a master package. Do you think that's any better? Also, curious if you have an opinion on the T-SQL approach. It's slower but I've tested it and it will work.
Dec
14
comment ETL: extracting from 200 tables - SSIS data flow or custom T-SQL?
The databases reside on multiple servers which I think makes it more complicated. I have actually tried creating a view of different tables on my development box against 225 databases and reading of the data was painfully slow.
Oct
27
comment overhead for INSERTS using change tracking and/or change data capture (CDC)
You may find this link useful: databasejournal.com/features/mssql/article.php/3806001/…
Oct
25
comment Data archival strategy that handles schema changes?
Maybe substantial is the wrong word but the functionality is constantly evolving. It would not be unusual to see the addition of new tables and columns as well as modifications to existing ones.
Aug
29
comment Does surrogate key assignment for a fact table require that the source data has natural keys?
In this example, I think the rules are one order per one customer and one product (not realistic, I know).
Aug
29
comment Does surrogate key assignment for a fact table require that the source data has natural keys?
I think so. Can you elaborate on what that mapping table might look like based on the example I have?
Aug
24
comment How do you roll up header-level values on a fact table with line-item granularity?
The query makes sense. But does this mean that the roll up can only occur within a view of some sort? If you were to take the fact table and build a cube from it, it seems like the customer wouldn't be able to do these roll ups in a pivot table.
Aug
24
comment How do you roll up header-level values on a fact table with line-item granularity?
@Jamie: But if I want to do a browse query on the Product dimension (say, all Products who fall under a certain brand), this design means I'll need to somehow filter out all my "magic" products. Is that also correct? It seems like the problem can be solved by having two fact tables: one for header level and one for line-item level. But nobody seems to recommend this.
Aug
24
comment How do you roll up header-level values on a fact table with line-item granularity?
Would the "magic" Shipping product be a placeholder so that a line item with 0 quantity and 0 order amount could exist with the shipping amount? Does this mean you would then zero out the Shipping amount on the other line items?
Aug
7
comment Filtering data ordered by rowversion
How does the date field fit in? Can a row with a particular ID and Date be updated in the table? And if so, is the date also updated (like an additional timestamp?)
Aug
2
comment How to break out of error loop in transactional replication?
I agree. For the future, I've modified the replication procedures so the error is logged. But the agent's attempt to "keep trying" seems like the equivalent of a service broker poison message. I don't see what is gained from not stopping the service.
Jun
29
comment How does SQL Server choose an index key for a foreign key reference?
@AaronBertrand: here's why I was asking if it was a bug: on SQL Server 2005, the behavior always chooses the primary key. On 2008, it chooses the primary key the first time if the primary key if the primary key is clustered. The behavior appears consistent, which is why it seems weird. But from what you're saying, it sounds like nobody but the programmers will ever know.
Jun
29
comment How does SQL Server choose an index key for a foreign key reference?
It makes sense...but the behavior is only on SQL Server 2008 R2 (not 2005, it's a complicated mess how I know all this), so I'm still leaning towards "potential bug."
Jun
29
comment How does SQL Server choose an index key for a foreign key reference?
I may not be understanding, but if I create primary key on the table as a clustered index in the example above, it will not use the duplicate index key as a reference (and the duplicate can be deleted). This appears to be related to it being a heap first...
Jun
26
comment How to log error details when using using try/catch for dynamic SQL backup commands
You may want to see this: stackoverflow.com/questions/5966670/…
Jun
24
comment How can I check the integrity of disabled foreign key relationships for each individual table? (SQL Server 2008)
Will your tables have primary keys or unique keys that you can use as references for the foreign keys? You won't be able to build these disabled foreign keys (in your example) without them.
Jun
16
comment How to efficiently check EXISTS on multiple columns?
@Thomas: Welcome to Stack Exchange. Have watched your SQL Bits Extreme DW presentation about twenty times.
Apr
30
comment ETL: extraction strategy for 200 source databases
Thanks for the detailed answer. Do you have any thoughts on using replication as a low impact means of mirroring the source data?
Feb
17
comment Data warehouse: using day number in Epoch in a date dimension
Interesting. In the example in the book, the last column in the date dimension is a SQL Date Stamp. So if Epoch means the same thing, I'm not sure why both would be included. But if Epoch is a timestamp, then what would "Day Number in Epoch" represent?