March 14, 2013 at 3:15 am
Hi guys,
I just wanted to know whether anyone else has experienced this and if so what the solution was.
I have a package that as part of a dataflow aggregates data, grouping on 5 columns and summing a sixth. The aggregate is supposed to create a single record which counts the number of lines in the datapipe ( the columns being summed always has a value of 1).
It has never failed in development, which is a 32bit laptop but on occassions in deployment it creates multiple output rows. the data viewer shows no difference in the aggregated columns.
My hypothesis is that before the aggregate I have a derived column task that adds a datetime using getdate() and that for some reason this is calculating different values even though the display value (which is only down to the second) is the same.
I am running BIDS2005 32bit on a Win2008 64bit virtual server when this behaviour is observed but it also happens when running from command line.
Unfortunately the package is permanently running as it uses Konesans file watcher so using the Package Start time variable is not an option for me. Also I dont want to use a script task to set a variable as we have no facility to support the vb script in the organisation (I am a contractor)
March 14, 2013 at 3:56 am
You might be right. Why not try changing your getdate() accuracy to one minute to test your hypothesis?
The absence of evidence is not evidence of absence.
Martin Rees
You can lead a horse to water, but a pencil must be lead.
Stan Laurel
March 14, 2013 at 3:57 am
Phil Parkin (3/14/2013)
You might be right. Why not try changing your getdate() accuracy to one minute to test your hypothesis?
Or use MAX(Getdate()) rather than grouping by it.
The absence of evidence is not evidence of absence.
Martin Rees
You can lead a horse to water, but a pencil must be lead.
Stan Laurel
March 14, 2013 at 7:44 am
Phil,
As usual, a simple solution to an annoying quirk of SSIS - Thanks. In this instance I am acutally only interested in the date so I have done a (DT_DBDATE)getdate(). I suppose I might still get a split if someone loads a file at 3 seconds to midnight...
Aaron
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply