Loading a whole lookup data file from disk into memory once, and then retrieving records from memory, is much faster than locating and retrieving records individually and repeatedly from a file stored on disk.
Data lookup: An analogy from the world of papers, folders, and cabinets
Components - Output file, Write Lookup, write multiple lookups
Lookup file component - Data file, key and record format
If a LOOKUP FILE component is not referenced by a DML core lookup, lookup_count, or lookup_next function in your graph, the component remains unopened, and lookup data is not loaded into memory during graph execution.
Ex - lookup("MyLookupFile", in.key)
Equivalent of an inner join or a left-outer join only
-
When to consider joining instead of lookup-plus-reformat
You need to perform a full outer join or a semi-join
Lookup data is huge which cannot be held up in memory