Merge stage : The parallel job stage that combines data setslookup stage: The mainframe processing jobs and parallel active stages that perform table lookups. Lookup stage: 1. Used to perform lookups.
2. Multiple reference links, single input link, single output link, single rejects link, single primary link. 3. Large amount of memory usage. Because paging required 4. Data on input links or reference links need NOT to be sorted. Merge stage:1. Combines the sorted data sets with the update datasets. 2. Several reject links, multiple output links will be exist. 3. Less memory usage.4. Data need to be sorted.
Both stages use for joining the data. using merger you can capture rejected data from N no of update datasets but you can't capture master source data. but using lookup you can capture 1 reject data set from source only not from the reference datasets.