with the iterate table, we can operate transformations on a reference table.
Could we have the same with repeat.
To reproduce a ‘with recursive’ SQL query in EM, we currently need to save the reference table as a dset file and import it into the repeated module.
the possibility to transfer a table would make process cleaner and maybe improve performance as with the iterate table transformation.
In the case of the regular “Repeat” action, the Input action of the iterated module receives either the input dataset of the “Repeat” action itself (in the 1st iteration), or the result dataset of the previous iteration (for the 2nd and consequent iterations).
In case the the suggested “Repeat table” action, the Input action would always receive the auxiliary table, which means that recursive iteration won’t be possible. Is recursivity required to reproduce a “with recursive” SQL query?
Here is an example.
in a loop I apply always the same table upon the result of the previous loop.
Here it allows me to start a reccursive exploration from a specific point instead of building an exploration from any point.
If I do it with a self join on the results of the previous loop, everything would be in relative path from any point.
EM repeat with table.zip (6.1 KB)
if we could even provide several reference tables.
the same with the iterate table transformation.