Question

Lets say I have 2 datasets: A and B, that have the same columns. I want to get the ‘distance path’ – minimal operations needed to be done, to change datasets A to B.

I can assume, that A and B contain only in few differences and I’m looking for something better, then:

• remove all A’s elements
• insert all B’s elements into A

How can this be archived?

Answer 1

If you want to find out how to transform A into B, you need to know what the difference between A and B is. There’s a very simple general algorithm for that, though it might be more complicated depending on how many fields are in your dataset and how many of them can be different from A to B. But here’s the general idea:

• Sort both datasets. Make sure to sort on the exact same criteria. (At this point, if they both contained the same set of data elements, the two lists of their contents would be identical if viewed side by side.)
• Begin with the first item of both
  datasets. Compare the two items with a comparison that can return less than, equal or greater than.
• • If A < B, note the difference and move to the next row of A.
• • If A > B, note the difference and move to the next row of B.
• • If A = B, advance both datasets to the next row.
• Repeat the compare-and-advance step until you’ve reached EOF on both datasets.

Once you’re done, you’ll have a full list of differences between A and B, which will make the steps needed to transform A to B simple to calculate.