No problem, really. I am/was going to run some timing tests on using a filter versus the dbpipeline begin/end record. I have noticed that some databases take a long time to open when they are filtered.
With the OnFilterRecord event, filters can take a long time if not all the data is already downloaded from the server. I have also seen some datasets that setting the 'Filter' property resends the request to the server for the data--which is needless extra processing (and if it is a complex record from multiple datasets can be rather expensive waiste of resources).
when using filters ensure that there is an index to speed it up. This should work for most databases. If there isn't a way to get the performance you need for a single record then I would create a 'JITPipeline'.
Comments
Yes.
Ed Dressel
BTW, if you want to know how (you didn't ask) look at the properties for
the pipeline with 'RageBegin' and 'RangeEnd'
what is your problem with using a filter?
regards,
Chris Ueberall;
filter versus the dbpipeline begin/end record. I have noticed that some
databases take a long time to open when they are filtered.
With the OnFilterRecord event, filters can take a long time if not all the
data is already downloaded from the server. I have also seen some datasets
that setting the 'Filter' property resends the request to the server for the
data--which is needless extra processing (and if it is a complex record from
multiple datasets can be rather expensive waiste of resources).
--
Ed Dressel
Team DM
when using filters ensure that there is an index to speed it up. This should work for most databases.
If there isn't a way to get the performance you need for a single record then I would create a 'JITPipeline'.
regards,
Chris Ueberall;
I won't call that a 'filter' and yes you are right using this event is mostly not performant.
Cheers,
Chris Ueberall;