Split list by keyword to individual lists.Or something??

To break it down into the main challenges:

The first step would be to assign all text line rows an Attribute value with the Employee Number.  This can be done with AttributeCreator run in Adjacent Features mode.

What we can do is get AttributeCreator to step through each text line and whenever the text line starts with “Employee:”  then this is the row that has the Employee Code value in it.  We can then use RegEx to extract it and with then we can keep assigning this code value to all subsequent rows until we reach the next line that has “Employee:” in it, which is where the next employee’s group of data starts.

 

@ReplaceRegularExpression(@Value(text_line_data),"^Employee:\s+(\d+).+$",\1)

The RegEx to extract the employee code puts the group of Digits “\d+” into RegEx string pattern Group 1 by putting the brackets “(...)” around it, and then we use this bracketed group of characters to use as the value for employee_number

Through the Adjacent Features we then assign this value to all subsequent rows except where we encounter a new row that again starts with “Employee:”

Output is:

From here is a Tester to just get the Code vs Value rows and get rid of the rest Eg.
 

Then we can use AttributeSplitter to create Lists per Employee per Code to capture the Unit values.  Assuming it is Tab delimited, then 

 

An important knowledge item in FME is that “Lists” are actually just normal Attributes but with an Attribute Name that is formatted like _list{0} , _list{1} etc.

We can use this knowledge for each row to Create New Attributes from each List Attribute using AttributeCreator to assign each List Attribute value to a “Normal” Attribute.