colunm resolving: uncleared resolving to a rename node

```yaml
- id: 4
mode: batch-request-unsupport
desc: LASTJOIN(SimpleOps(FILTER))
sql: |
SELECT
t1.c1,
t1.c2,
t2.c1 as c21
FROM t1 last join (select c1 as c1r, c2, c3, c4 from t2 where c1 != 'aa') t2
on t2.c1r = t1.c1;
inputs:
- name: t1
columns: ["c1 string","c2 int","c3 bigint","c4 timestamp"]
indexs: ["index1:c1:c4"]
rows:
- ["aa",2,3,1590738989000]
- ["bb",21,31,1590738990000]
- ["cc",41,51,1590738991000]
- name: t2
columns: ["c1 string","c2 int","c3 bigint","c4 timestamp"]
indexs: ["index1:c1:c4"]
rows:
- ["aa",2,13,1590738989000]
- ["bb",21,131,1590738990000]
- ["dd",41,151,1590738991000]
expect:
columns:
- c1 string
- c2 int
- c21 string
order: c1
data: |
aa, 2, NULL
bb, 21, bb
cc, 41, NULL
```

`t2.c1 as c12` should not work ? It should be `t2.c1r as c12`

The issue is about column resolution in SQL queries with renamed columns, specifically in a LAST JOIN context. Look at the SQL parsing and column resolution logic in the codebase, likely in the query planner or optimizer. The test case provided is a YAML specification; find where such tests are run to understand the expected behavior. The fix involves ensuring that column references after a rename (like `c1r`) are correctly resolved.

Scritto dal modello di indicizzazione a partire dal testo della issue.