Currently we use a form of depth-first-search to detect (and then cut out) all of the nodes (tables) that are part of OR lead to a cycle.
We should use an algorithm like Tarjan's Algorithm that detects the strongly connected components inside the graph and only cut those out.
This will lead to sql-scripts that have less "alter-table"-Statements at the end of them and therefore have increases readability
Currently we use a form of depth-first-search to detect (and then cut out) all of the nodes (tables) that are part of OR lead to a cycle.
We should use an algorithm like Tarjan's Algorithm that detects the strongly connected components inside the graph and only cut those out.
This will lead to sql-scripts that have less "alter-table"-Statements at the end of them and therefore have increases readability