This Snap eliminates duplicate documents in a document stream, such as duplicate rows in a CSV file. To understand the functionality of the Unique Snap, it is important to understand that SnapLogic is streaming data, processing "one document at a time" in each Snap. 

If the data being passed was in the following order:

 
It would recognize that the data was in pairs and the result would be just the 10 de-duplicated rows: 

 
If the data however were in a different order:  

The result would be exactly the same as the input, as the duplicate rows are not adjacent in the flow and therefore, are not identified.
 
If you wanted to eliminate the duplicates, it is possible to use a Sort Snap in the flow before the Unique, in which case the duplicates would be adjacent, and hence de-duplicated. 

• Expected upstream Snaps:  Any Reader followed by any file Parser.
• Expected downstream Snaps:  Any file Formatter followed by a Writer.
• Expected input:  Document input, likely with duplicate data.
• Expected output:  Unique document data.