group function does not give error when column name is not existed

this query just groups data by "shareholder_shareholder_id" and does not raise error for "shareholder_share_id" column that not existed anymore because we changed the name of it

```sh
from(bucket: "day")
|> range(start: v.timeRangeStart, stop: v.timeRangeStop)
|> filter(fn: (r) => r["_measurement"] == "shareholder")
|> pivot(
rowKey:["_time"],
columnKey: ["_measurement", "_field"],
valueColumn: "_value"
)
|> duplicate(column: "shareholder_share_z", as: "total_share_z")
|> rename(columns: {shareholder_share_id: "symbol_id"})
|> group(columns: ["shareholder_share_id","shareholder_shareholder_id"], mode:"by")
```

Start by running the Flux query in the issue and verify how group handles the renamed, nonexistent column. Trace the group operation from that reproduction to identify where missing columns are accepted; done means the query reports an error for the absent column while valid group columns continue to work.

Written by the indexing model from the issue text.