if the reference datasets are big enough , use a join. A join does a high-speed sort on the driving and reference datasets.Once the sort is over the join processing is very fast
join has no reject links. wheras lookup has one reject link join has 4 types inner outer,left outer,rght outer. l0ok up gives inner and left outer lokk up takes any kind of colmn names and perform join operations.join needs same coloumn names among all sources. if refernce table has little data prefer to lookup more means prefer join..
Lookup stage requires all the data to be loaded in memory 1st & then only it can start matching process with inputs.
Hence It is advisable to use join stage then Lookup stage.
If you want to capture Reject records then you can use transformer (or filter) after Join Stage and capture the null values. Or Instead of join stage use Merge Stage to capture the reject records.
