Semantic Clarification for Result of Multiple Selects

[awalley-ncqa]

Under the multiple selects section of the documentation it says that

"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:

  "resource": "Patient",
  "status": "active",
  "select": [
    {
      "column": [
        {
          "name": "id",
          "path": "id",
          "type": "id"
        }
      ]
    },
    {
      "column": [
        {
          "name": "last_name",
          "path": "name.family.first()",
          "type": "string"
        }
      ]
    }
  ]
}

and results in

id	last_name
pt1	F1
pt2	F2
pt3	null

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.

[arjun]

NR: probably bug in the FP engine being used. What should be behavior? Consult with John and Brian P.

[johngrimes]

I think this is covered off in the Processing Algorithm section, please re-open if there are any concerns.