You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When using ORDER BY 2, 1, users are referencing fields by index. If such a field is a nested record (#11812), or an embeddable record (#10527), then we must replace the index by an expanded row field, if the dialect has no native support. I.e.:
-- With native support, e.g. PostgreSQL:SELECT a, row(b, c), d
FROM t
ORDER BY2, 1, 3-- Without native supportSELECT a, b as"row.b", c as"row.c", d
FROM t
ORDER BY2, 3, 1, 4
In other words:
We should continue using field indexes, not expressions, etc. i.e. 2 -> 2, 3
We must expand individual indexes by their index list
We must shift indexes after a row reference, i.e. 3 -> 4
The text was updated successfully, but these errors were encountered:
When using
ORDER BY 2, 1
, users are referencing fields by index. If such a field is a nested record (#11812), or an embeddable record (#10527), then we must replace the index by an expanded row field, if the dialect has no native support. I.e.:In other words:
2
->2, 3
3
->4
The text was updated successfully, but these errors were encountered: