[BUG] DataFrame.update() should modify the memory inplace as opposed to replacing the buffers

[kkraus14]

In #6883 DataFrame.update() was implemented but it was implemented to replace the Columns underneath the DataFrame. In the Pandas implementation it actually mutates the memory underneath the DataFrame which can have propagating effects:

import pandas as pd

pdf = pd.DataFrame({'a': [1,2,3], 'b': [4,5,6]})
pdf2 = pd.DataFrame({'c': [7,8,9], 'b': [14,15,16]})

# before update
initial_pointer = pdf._data.blocks[0].values.__array_interface__['data']

pdf.update(pdf2)

# after update
final_pointer = pdf._data.blocks[0].values.__array_interface__['data']

assert(initial_pointer == final_pointer) # doesn't Throw

Not high priority but wanted to raise an issue to have something to point to in case this comes back.

[kkraus14]

cc @skirui-source @galipremsagar @isVoid for visibility

[github-actions]

This issue has been labeled inactive-30d due to no recent activity in the past 30 days. Please close this issue if no further response or action is needed. Otherwise, please respond with a comment indicating any updates or changes to the original issue and/or confirm this issue still needs to be addressed. This issue will be labeled inactive-90d if there is no activity in the next 60 days.