The Briggs & Torczon paper focuses on set representation rather than key-value style access, but of course one can get associative access with just a 2nd dense

The Briggs & Torczon paper focuses on set representation rather than key-value style access, but of course one can get associative access with just a 2nd dense array of satellite data or, equivalently, wider-than integers in the one dense array (depending upon memory layout desires). Then you basically have a direct access index (like only 256 possible things, so use an 8-bit byte-sized index array of 256 entries) paired with a dense table. This always struck me as very similar to early ideas in database indexing a la Ted Codd & crew (as perhaps more recently popularized in recent Python3.6 hash table impls). So, while Aho & Briggs & Torczon all deserve ongoing credit for popularization, I wonder if anyone out in HN-land knows of a reference to a database paper from the 60s or 70s introducing this? (For more complete attribution.)