SQL Server Performance Forum – Threads Archive
End to end testing dw processes
Hello people. End to end testing dw functionality is problem we don’t have satisfactory solution yet. We have daily dts packages feeding dw from production. After data are integrated some application specific data marts are refreshed and some summary data are sent back to production dbs. Because of huge amount of data and complex transformation and overall processes it is dificult to define efficient qa process for testing if information flow from production dbs trough dw and back to the production (and/or applications data mart) is correct. I would like to learn about experiance other people have with this.Hi, I am years from calling myself a semi-experienced person in this area but here’s sharing my experience. Ihope this helps to some extent? I had the someworries a few monthsback becuase though the overall totals across 5 dimension and 6 measures were tallying with the results of the OLTP there were still millions of possibilities. We have been doing a manual testing on a daily basis (with additional manpower) and have found a few shortcuts e.g. exporting the results fromthe front-end into Excel ina certain manner and order and having identical data from OLTP in Excel and then pasting them next to oneanaother and simply putting ‘= CellA – CellB’in a column named "check" and then using thefilter feature of Excel to see if there are any records where the Check column is not equal to zero—> this would represent danger but luckily no such encounter. I don’t know if this helps but please do share your experience (and Iwould request this from other experienced users too) and really would like to know if there is a better/automated way of doing this. Regards.
]]>