On the GROUP_ARRAY_INTERSECT page (aggregate-functions/group-array-intersect.md, version-3.x and version-2.1), the first example does not match its own output:
select c_array_string from group_array_intersect_test where id in (18, 20);
+------+---------------------------+
| id | col |
+------+---------------------------+
| 1 | ["a", "b", "c", "d", "e"] |
| 2 | ["a", "b"] |
| 3 | ["a", null] |
+------+---------------------------+
The SQL selects a single column c_array_string with predicate id IN (18, 20), but the shown output has columns id | col and rows with id = 1, 2, 3 (which contradicts the IN (18, 20) filter). The SQL and the displayed result are inconsistent.
Suggested fix — make the SQL match the shown two-column output and rows, e.g.:
select id, col from group_array_intersect_test;
(The second example, select group_array_intersect(col) from group_array_intersect_test; → ["a"], is consistent and fine.)
Filing this separately so the example-SQL correction can be reviewed on its own; the missing CREATE TABLE setup for these pages is being handled in the Line-B setup PRs.
On the
GROUP_ARRAY_INTERSECTpage (aggregate-functions/group-array-intersect.md, version-3.x and version-2.1), the first example does not match its own output:The SQL selects a single column
c_array_stringwith predicateid IN (18, 20), but the shown output has columnsid | coland rows withid= 1, 2, 3 (which contradicts theIN (18, 20)filter). The SQL and the displayed result are inconsistent.Suggested fix — make the SQL match the shown two-column output and rows, e.g.:
(The second example,
select group_array_intersect(col) from group_array_intersect_test;→["a"], is consistent and fine.)Filing this separately so the example-SQL correction can be reviewed on its own; the missing
CREATE TABLEsetup for these pages is being handled in the Line-B setup PRs.