04-27-2016 02:40 AM
When using a lookup: does the columns selected in the lookup table impact the performance?
For example: columns in LookupTable: ID, Name, Adress, Telephone, Age, Gender, ... and 200 other columns.
When doing a lookup on ID, i only want to retrieve the Age.
Will it matter that my other columns are available in the lookuptable or not? Or do i have to filter out all other columns before the lookupstep?
04-27-2016 10:39 AM
Having data sets with 200 columns sounds like your data is not modelled properly...
Sure, you could drop columns. But if you need them later on, any join operation will cost much more than gain to drop them.
The Lookup transform uses a data step that reads the source data sequentially - which usually is not very expensive. What drives performance in the Lookup is the no of lookup tables, and their size.