SSIS 2008 or 2012 Data Task: How to conditionally choose one of two flatfile columns on a per-row basis

I searched the textbooks and searchable forum posts for something related to my question, but could not find it. I'm hoping when after I describe my question that it might sound familiar to someone.

First, the context: another department using a black-box database product stopped use of one column and used a different column instead. However, the data in the flat-file export that I need to use for my own database must contain the entire table. If it was a clean binary switch, that would be great: I could split the rows into two groups based on date and deal with it that way. Instead, there was a period of overlap while users slowly got the hang of the practice.  This means that I have two columns where a row value is either in one columnor the other.

Now, the question: I currently use SSIS to import this flatfile into an SQL Server table and use a T-SQL script in an SQL Agent Job to handle the problem.However, I was hoping I could do this on the SSIS side while importing the data so that it appears into the SQL Server destination table nice and clean.

This is a HIGHLY SIMPLIFIED view of my system, using dummy data and dummy names. The T-SQL script describes in essence my source and destination table and what I would like to do on the SSIS side. If a screenshot of my incomplete data tab will help, I can do it, but at this point I'm providing what I hope presents the background without presenting distractions.  

Thank you for any hints or ideas.

use Learning_Curve;
go

if OBJECT_ID('dbo.Relation_A','U') is not null
	drop table dbo.Relation_A;

create table dbo.Relation_A (
	Rel_A_ID int primary key,
	Rel_A_Value varchar(20) not null,
	Rel_B_Value varchar(20) not null
);

insert into dbo.Relation_A values (1,'Unknown','Measured'),(2,'Measured','N/A'),(3,'Unknown','Measured'),
								  (4,'Measured','Unknown'),(5,'N/A','Measured'),(6,'Unknown','Measured');

if OBJECT_ID('dbo.Relation_D','U') is not null
	drop table dbo.Relation_D;

create table dbo.Relation_D (
	Rel_D_ID int primary key,
	Rel_D_Value varchar(20) not null
);

insert into dbo.Relation_D
select
	a.Rel_A_ID as Rel_D_ID,
	case
		when a.Rel_A_Value = 'N/A' or a.Rel_A_Value='Unknown'
		then a.Rel_B_Value
		else a.Rel_A_Value
	end as Rel_D_Value

	from dbo.Relation_A a;
go