You could set up a choice parameter to allow the user to pick one (just one) of the options. A Creator to create a dummy feature and a TestFilter to route that feature based on the parameter value. Then 2 seperate FeatureReaders to do the actual data reading.
That works great, thank you - but do you know how you can send more than one choice, i.e. select both sets of data and process one before processing the second, and also selecting more than one writer??? Thanks again.
Not quite an answer to your follow up question, more an elaboration to my initial answer that I had to write up for a customer who filed a similar support question yesterday. I thought it would be useful for more people so I'm adding it here.
Their situation: it's 3 cities cooperating with eachother. Their databases are seperate, but the schemas are identical. The joint geo department can access all 3 databases, so they have 3 connections set up in FME and wanted to know how they can easily allow a user to select one of the three cities and process the data without setting up every reader three times. This is done through a SQL query with a Where clause including a city code.
I've mimicked their setup by creating 3 PostGIS databases (Bar-test-1 through -3) with some sample data (all the same) and 3 connections in FME, named Bar-test-1 through -3 as well.
Step 1 is to set up a Choice with Alias parameter to map the city names to city codes
Step 2 is to set up a Creator and an AttributeManager to create a dummy feature that will kick things off.
Step 3 is to set up the AttributeManager to add an attribute called "connectie" (connection) with a conditional value based on the parameter
Step 4 is a SQLExecutor to fire off the query to the correct database. The connection name is taken from the attribute and in the query itself the parameter is used in the Where clause.
And presto!