Pandas dataframe short traceback

[lainisourgod]

Hi! I had a problem for a while that rich traceback with show_locals=True tended to produce reeeealy long tracebacks because of dataframe printing that printed 10+ wide columns with wraping them on multiple lines, and that is for 10+ rows of dataframe, and 10+ frames of traceback. So it was really annoying to look through traceback.

The solution I found is: pd.set_option("display.max_columns", 3). This way traceback is pretty lean, and I don't really miss info because I rarely need to debug the whole datarame content.

Code to reproduce:

import pandas as pd
from rich.traceback import install

install(show_locals=True)

# NOTE: This is the option to reduce traceback length!
# pd.set_option("display.max_columns", 3)

big_df = pd.DataFrame(
    {
        "some_long_feature_name": [1] * 100,
        "another_long_feature_name": "hello" * 100,
        "even_longer_feature_name_with_wide_data": [pd.to_datetime("2023-01-19")] * 100,
        "even_longer_feature_name_with_wide_data_1": [pd.to_datetime("2023-01-19")]
        * 100,
        "even_longer_feature_name_with_wide_data_2": [pd.to_datetime("2023-01-19")]
        * 100,
        "even_longer_feature_name_with_wide_data_3": [pd.to_datetime("2023-01-19")]
        * 100,
        "even_longer_feature_name_with_wide_data_4": [pd.to_datetime("2023-01-19")]
        * 100,
        "even_longer_feature_name_with_wide_data_5": [pd.to_datetime("2023-01-19")]
        * 100,
        "even_longer_feature_name_with_wide_data_6": [pd.to_datetime("2023-01-19")]
        * 100,
    }
)


def fun_1(df: pd.DataFrame):
    fun_2(df)


def fun_2(df: pd.DataFrame):
    fun_3(df)


def fun_3(df: pd.DataFrame):
    raise asfs


fun_1(big_df)

[lainisourgod]

Also, just realized you can also set pd.set_option("display.max_rows", 3) to minimize traceback even more