Which way is better for storing XML messages as test samples?

Here is the scenario:

I have developed an API-test case which needs 10 different xml-message samples to iterate over them. I am confined to use relational DB for storing the samples. The XML messages contain too many elements such that I cannot define a table for each message, but on the other hand I don't have to examine all the elements(just a few important ones!). Two different solutions came across my mind:

1) Create a table with those important elements(fields) and construct the xml-message samples dynamically by retrieving these values from DB. In this case I need to add a new table for each kind of xml-messages.

2) In my TEST_CASES table, I can store whole xml-message samples in BLOB type and then for every iteration of the test case I just need to decompress the message and feed it to the test flow. But in this case I cannot easily manipulate the samples by changing the corresponding fields in the table like first solution.

If you have any better ideas I would appreciate your suggestion, otherwise let me know which one of those mentioned approaches you would pick.