RESTORE job description is sometimes semantically incorrect

**Describe the problem**

The two following two restores:

```
> select job_id, job_type, description, statement from [show jobs] where job_type = 'RESTORE';
job_id | job_type | description | statement
---------------------+----------+----------------------------------------------------------------------------------------------------------------------------+------------
948579866317520897 | RESTORE | RESTORE DATABASE defaultdb FROM 'nodelocal://1/restore-desc/2024/03/04-120551.27' WITH OPTIONS (new_db_name = 'restore_1') |
948580038300106753 | RESTORE | RESTORE DATABASE defaultdb FROM 'nodelocal://1/restore-desc/2024/03/04-120551.27' WITH OPTIONS (new_db_name = 'restore_2') |
```

Look very similar with the exception of the db name. But, they are actually from very different restore attempts and restored different data:

```
demo@127.0.0.1:26257/demoapp/defaultdb> select * from restore_1.foo;
pk
------
1
2
3
4
(4 rows)

Time: 2ms total (execution 1ms / network 0ms)

demo@127.0.0.1:26257/demoapp/defaultdb> select * from restore_2.foo;
pk
------
1
2
(2 rows)
```

**To Reproduce**

- Create a backup with a few incremental layers.
- Restore from the latest incremental layer
- `restore database defaultdb from latest in 'nodelocal://1/restore-desc' with new_db_name='restore_1'; `
- Take the description from that job and use it to restore again.
- Note that both have the same job description but the second job will be a restore from the first full layer, not the latest incremental layer.

Ideally, we would preserve the collection-based `IN` syntax when it was used originally.

Jira issue: CRDB-36361