ENH: context-manager for chunksize/iterator-reader

[twoertwein]

• closes #xxxx
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

Allows (and encourages) the following use:

import pandas as pd

filename = "pandas/tests/io/data/csv/iris.csv"
chunksize = 2
with pd.read_csv(filename, chunksize=chunksize) as reader:
    for chunk in reader:
        # risky code that might raise

Same can be done for read_json/sas (I think these are all methods that support chunksize). If this PR should make it into 1.2, I can quickly add the changes for json/sas as well.

Are there more places to promote this new context manager?

[jreback]

this looks pretty neat. can you add a small note section in io.rst which shows this off. i think this is ok for 1.2, can add json/sas as followsup.

[twoertwein]

I don't understand why the documentation example is failing:

Exception in /home/runner/work/pandas/pandas/doc/source/user_guide/io.rst at block ending on line 1586
ParserError: Error tokenizing data. C error: out of memory

edit: the blank line was causing the issue, it was interpreted as the end of the with-block

[jreback]

great. just some minor doc omments. ping on green.

[twoertwein]

@jreback green. I hope the whatsnew entry is good now

[jreback]

thanks @twoertwein very nice