You have a lot of options here but which one you choose will depend on your use case. I would possibly load the data into Stardog, either directly or via virtual mapping, and serve the tables at some IRI with a web server and include a link to the html table with your data. You'd have the data that your can search against and when you wanted the html table for presenting you'd have them fetch from the IRI.
There's a lot of cool stuff to checkout with Stardog.
One thing I'd add to my answer that just occured to me. This is assuming that your tables are generally immutable since keeping any updates synched between your HTML tables and the triples stored in stardog difficult. You could also write a custom BITES extractor that would extract the table data automatically and provide and HTTP endpoint for the tables automatically. The data would be updated when you reprocessed the table in BITES. You could also write some custom code to simply recreate the HTML tables directly from triples with something like an XSLT to transform a sparql results query to an HTML table.
You have a bunch of options here depending what exactly you're looking to do.