I've never used the feature, but my understnading is that it only works with images -- you need metadata for the caption to read.
Why not just add a text frame, or a new paragraph below the table in the same frame?
Yes, I realize that Adobe has crafted the Caption function to import the static name of the graphics file, or metadata. However, if you think about it from a publishing perspective, shouldn't tables have captions, too? The Caption function should insert an appropriately sized and grouped text frame at the bottom, with automatically applied object and paragraph styles both named Caption unless overwritten by the user. Static and Dynamic caption text should be options for this, but "fill in the blank" should be an option, too. Stupid design. As Alan Cooper would say, this is an example of user interaction design based on an implementation model (how the programming code modules are organized), and not the user's mental model. The user interface is modeled after what the coders designed, not what users expect. And if you want tables in your book in addition to graphics, you're out of luck if you try to use the Caption feature.
Why not use a footer row for that?