jupyter notebook display bug for pandas dataframes with double indexes

Summary

Dataframes with double indexes displayed in jupyter notebooks can be misaligned under certain conditions.

Steps to reproduce

Create a jupyter notebook outputting a pandas dataframe where the dataframe has two indexes and the primary index contains duplicate entries.

Example Project

https://gitlab.com/DeltiKron/double-index-jupyter-display

What is the current bug behavior?

The primary index cell is omitted in rows where the primary index is identical to the previous row, making these rows one cell shorter and throwing the rows out of alignment.

What is the expected correct behavior?

The corresponding rows should be left-padded with an empty cell.

Relevant logs and/or screenshots

Expected behavior as seen in a native jupyter notebook:

Behavior on GitLab.com:

Output of checks

This bug happens on GitLab.com