What’s Special About Ancelus Concatenated Keys?

The example above shows how a pricing is handled where more than one vendor exists for each part, and more than one part is supplied by each vendor. The double headed arrow indicates a one-to-many relationship. If we use only the two primary entities, parts and suppliers, the requirement is for a many-to-many relationship which is logically un-resolvable. To break this log-jam we create a new entity call the PN-VEN representing the concatenation of part number and vendor. This now has a unique resolution to price. There are other items that might be handled in a similar way, for example, lead time.

Concatenated keys are common in any database that needs to index on the combination of more than one key field. In relational systems the method is to create a table with each of the key fields, then index on the combination. The data in the new table is a copy of the original source data (which now exists in two places).

Ancelus does the same thing, but the mechanics are different. Assume we are interested in three 40-byte key fields, the concatenated table will appear to be 120 bytes wide (plus the index). In Ancelus that actual table will only be 12 bytes wide (plus the index), since the Ancelus table is virtual and contains no actual data. It contains pointers to the actual data and in the 4 series system, the pointer is 4 bytes. The data resides elsewhere in the system and physically exists only one time. The index for the concatenated key is integrated and automated in that it is automatically updated any time a new record is entered into the concatenation. The same applies for the deletion of a record. The index is updated and the storage space is immediately made available for re-use.

The more complex the concatenation, the larger the compression ratio. In our workflow automation system, the event record looks like it has 680 bytes, but the actual size is only 40 bytes. The entire system operates as a 7-dimensional hypercube, operating in real-time in all 7 dimensions.

This is one of the examples of how Ancelus compresses the size of the dataset and simplifies the management of the indexes.   

To learn more or to discuss your specific design, contact us at www.ancelus.com.