Actually Using my First Model?

WindsAloft

I've successfully build my first model, which is able to classify/categorize records in a database based on text mining the comments field.  My model gives me a high degree of accuracy - but how do I actually implement this to future (unknown category) records? 

Can I save the model so that it doesn't have to re-read the training set every single time before trying to predict?  Also on my output, I'm unable to pass the primary key from the source to the results - obviously the primary key isn't used in the algorithm so it whines when I try to pass through a column that isn't actually used anywhere in the model.

Perhaps if theres a single example I could look at - that shows end to end how you can use a dataset to train/validate and then pass a new, completely unknown dataset through the process and output the original dataset (all column) + the predicted column.