[Template-haskell] is this (.csv) a good task for template-haskell?

[Lloyd Allison]

I'd like to write some Haskell programs to manipulate
comma-separated variable, .csv, files from a spreadsheet.

The obvious way to start which, for perverse(?) reasons,
I do not want to follow is to create a union type e.g.
  data CSVelement = I int | S string | ...
or similar.

All the elements of a given column will be of the same type
 e.g. bool, int, float,
      string, or
      a-selection-of-names (e.g. {Lab, Con, LibDem}).
(and some columns with missing values).

e.g. "Homer",42,67.8,Lab
     "Paris",,5.5,Con
     ...

I'd like to write some code that
 1. takes a quick look at the file,
    counts the columns,
    works out the type of each column
    (including creating data T1 = Lab|Con|LibDem,
    and using `Maybe' for missing values),
 2. creates a tuple-type for a row,
    e.g. (String, Maybe Int, Float, T1), and
    adds that type to Reads, Show (and some other stuff), and
 3. then gets on with processing the data.

Is template-Haskell a good choice for this task and,
if so, what is the quickest way to get started?
(Or has it been done?)

-Lloyd
--
Lloyd ALLISON,  CSSE, Monash University, Victoria, Australia 3800.
web: http://www.csse.monash.edu.au/~lloyd/    tel: +61 3 9905 5205