Question: `preprocess_xgboost`

I have a question about [BQML notebook](https://github.com/GoogleCloudPlatform/training-data-analyst/blob/master/quests/vertex-ai/vertex-bqml/lab_exercise.ipynb).
In the double for loop at the function `preprocess_xgboost`, `instance_list` seems to be overwriten. Is it correct?(`instance_list = list(instance.values()) `should be aligned with `transformed_instances.append(instance_list)`?)

```
def preprocess_xgboost(instances, cat_feature_list, feature_encoders):
"""Transform instances to numerical values for inference.
Args:
instances (list[dict]): A list of feature dictionaries with the format feature: value.
cat_feature_list (list): A list of string feature names.
feature_encoders (dict): A dictionary with the format feature: feature_encoder.
Returns:
transformed_instances (list[list]): A list of lists containing numerical feature values needed
for Vertex XGBoost inference.
"""
transformed_instances = []

for instance in instances:
for feature in cat_feature_list:
feature_int = feature_encoders[feature].transform([[instance[feature]]]).item()
instance[feature] = feature_int
instance_list = list(instance.values())
transformed_instances.append(instance_list)

return transformed_instances
```