Adjust plans for new queries in privileges tests

The upstream privileges regression test added multiple checks of explain
plans, so the plans needed to be adjusted for Postgres-XL (by adding the
Remote Subquery nodes to appropriate places).

There are two plans that however mismatch the upstream version, using
a different join algorithm (Nested Loop vs. Hash Join). Turns out this
happens due to Postgres-XL not collecting stats for expression indexes,
and the two queries rely on that feature. Without the statistics the
estimates change dramatically, triggering a plan change.

We need to extend analyze_rel_coordinator() to collect stats not only
for the table, but for all indexes too. But that's really a matter for
a separate commit.

0 files changed, 0 insertions, 0 deletions