import numpy as np import pandas as pd n_rows = 1_000 group_size = 10 n_random_cols = 200 data = {"id": np.repeat(np.arange(n_rows // group_size), group_size)} for i in range(n_random_cols): data[f"col_{i}"] = np.random.randn(n_rows) df = pd.DataFrame(data) # PerformanceWarning when as_index is False named_agg_without_index_warning_df = ( df .groupby('id', as_index=False) .agg(**{ column: pd.NamedAgg(column=column, aggfunc="mean") for column in df.columns if column != "id" }) ) # no warnings when as_index is True named_agg_with_index_ok_df = ( df .groupby('id', as_index=True) .agg(**{ column: pd.NamedAgg(column=column, aggfunc="mean") for column in df.columns if column != "id" }) ) # no warnings when using dict agg no matter what as_index is dict_agg_ok_df = ( df .groupby('id', as_index=False) .agg({ column: "mean" for column in df.columns if column != "id" }) )