"Sibling selects are effectively cross joined, where each row in each select is duplicated for every row in sibling selects."
This leads me to believe that when you have two sibling selects, a data table should be created for each select (from every resource) then the two of those are CROSS JOIN'ed i.e. cartesian product'ed.
This test example is taken from basic -> two selects with columns:
With the wording above, I would expect the result to be
id
last_name
pt1
F1
pt1
F2
pt1
null
pt2
F1
pt2
F2
pt2
null
pt3
F1
pt3
F2
pt3
null
I believe that the wording could be re-worked to say something along the lines of "for each resource, the data tables resulting from sibling selects are CROSS JOIN'ed. Then the rows from the cross join are union'ed into the final data table" if the desired result in the test is correct.
Under the multiple selects section of the documentation it says that
This leads me to believe that when you have two sibling selects, a data table should be created for each select (from every resource) then the two of those are CROSS JOIN'ed i.e. cartesian product'ed.
This test example is taken from basic ->
two selects with columns
:and results in
With the wording above, I would expect the result to be
I believe that the wording could be re-worked to say something along the lines of "for each resource, the data tables resulting from sibling selects are CROSS JOIN'ed. Then the rows from the cross join are union'ed into the final data table" if the desired result in the test is correct.