Mapping is the process of performing transformations to 
resolve data representation differences between services or 
document formats. By linking variables to each other on the 
Pipeline tab, you can accomplish name transformations and 
structural transformations. However, to perform value 
transformations you must execute some code or logic. 
Developer provides two ways for you to invoke services: You 
can insert INVOKE steps or you can insert transformers onto 
the Pipeline tab. Transformers are the services you use to 
accomplish value transformations on the Pipeline tab.

When we are performing only one data operation we can go 
ahead with directly invoking the corresponding flow 
service. If we want to perform the multiple operations in a 
single step, it is good to go with transformers in a map 
step. This increases the memory usage and readability.

Transformer emcompass mapping logic. Flow Service can be 
implemented as a trsndsformer.

Both invoking Flow services and 
transformers can be used to provide the same functionality,
but there are certain differences
In case of transformers the order of execution 
is not strict(i.e the order of transformer execution)
is not determined and multiple transformers can be 
executed at the same time.
And also transformers cannot be used whenever the
underlying data structure dimension is different(for ex: 
can't execute a transformer on a document when iterating 
over a loop on a document list)

Services that you insert using the INVOKE step might also 
perform value transformations. However, only transformers 
can accomplish multiple value transformations in a single 
flow step.