How can I compress similar rows in SQLite?

I'm using SQLite (via System.Data.SQLite if that matters) to story a very large number of blobs of varying size (~10-100kb) in a key-value fashion. Rows are appended, never deleted.

Since many of the blobs are very similar, it would be possible to compress them to a high degree if they were concatenated. However, that's not trivial to apply here, given that the data is in different rows.

Is there a reasonably easy way to still compress them, or at least remove part of the redundancy?