Expected behaviour for sort queries

Continuing the discussion from Identical query execution has inconsistent response:

This does not seem like the right behaviour! No other database does this that I know of. I am familiar with SQL and MySQL so from that realm, if I sort based on nullish field in a column it will put the nullish order on top. This can be controlled with using CASE in the ORDER BY clause to move the nullish values to the end of the order though.

Dgraph should have some kind of equivalent sorting of missing/nullish predicates. I understand that null predicates in Dgraph === missing predicates, which is not exactly the case in RDBs

Another Related Post:

1 Like