Redundant outer DISTINCT adds Sort and Unique above INTERSECT

[陈列行]

## Description

This issue concerns an outer `DISTINCT` applied to an `INTERSECT` result.
Since non-`ALL` `INTERSECT` already returns duplicate-free rows, the outer
operation cannot change the result. PostgreSQL nevertheless performs a
second deduplication.

### Expected Behaviour

PostgreSQL should propagate the uniqueness guarantee from `INTERSECT` and
remove the outer `DISTINCT`. Both equivalent forms should use the same plan
and have comparable execution times.

### Actual Behaviour

The outer `DISTINCT` adds `Sort -> Unique` above `HashSetOp Intersect`. With
two 500,000-row inputs and 250,000 result rows, five-run median execution
time increased from 176.081 ms to 191.225 ms, approximately 8.6%.

## How to repeat

```sql
DROP TABLE IF EXISTS intersect_distinct_lhs;
DROP TABLE IF EXISTS intersect_distinct_rhs;

CREATE TABLE intersect_distinct_lhs (v INTEGER NOT NULL);
CREATE TABLE intersect_distinct_rhs (v INTEGER NOT NULL);

INSERT INTO intersect_distinct_lhs
SELECT g FROM generate_series(1, 500000) AS g;

INSERT INTO intersect_distinct_rhs
SELECT g FROM generate_series(250001, 750000) AS g;

ANALYZE intersect_distinct_lhs;
ANALYZE intersect_distinct_rhs;

EXPLAIN (ANALYZE, BUFFERS, VERBOSE, SETTINGS, TIMING OFF)
SELECT DISTINCT v
FROM (
    SELECT v FROM intersect_distinct_lhs
    INTERSECT
    SELECT v FROM intersect_distinct_rhs
) AS set_result;

EXPLAIN (ANALYZE, BUFFERS, VERBOSE, SETTINGS, TIMING OFF)
SELECT v
FROM (
    SELECT v FROM intersect_distinct_lhs
    INTERSECT
    SELECT v FROM intersect_distinct_rhs
) AS set_result;
```

Both queries return the same 250,000 rows. The characteristic plans are:

```text
with DISTINCT: Unique -> Sort -> HashSetOp Intersect
without:       HashSetOp Intersect
```