BUG: options date_dayfirst and date_yearfirst conflict

[attack68]

Can someone confirm that this is a bug:

pd.options.display.chop_threshold = 0.5
pd.options.display.date_dayfirst = True
df = pd.DataFrame([[1.00, 0.1, 0.01, 0.001], [2.22, 2.22, 2.22, 2.22]], 
                  index=pd.date_range(start="01-01-2000", periods=2))
print(df)
               0     1     2     3
2000-01-01  1.00  0.00  0.00  0.00
2000-01-02  2.22  2.22  2.22  2.22

Also it seems weird to have the options display.date_dayfirst and display.date_yearfirst with both set to False by default, and yet print year first..

Docs say: When True, prints and parses dates with the day first, eg 20/01/2005. Maybe this is just for parsing?

[k4thir]

Hi , I'm a beginner to open source and can I contribute on this ?

[attack68]

@k4thir not really, because the error has not been triaged or discussed, so it is unclear what you should do. Instead trying searching for pandas issues with label:"good first issue" and milestone contributions welcome.

[JARVIS-03]

Hi @attack68,

To set the date format for your dataframe index, you have to use strftime(format)

Please add the following line in your code to get the index date to start from the day

df.index = df.index.strftime('%d/%m/%Y') # 01/01/2000

Some of the Date format available,

df.index = df.index.strftime('%d/%m/%Y') # 01/01/2000
df.index = df.index.strftime('%m/%d/%Y') # 01/01/2000
df.index = df.index.strftime('%Y/%m/%d') # 2000/01/01

For more info, please go through the below link,
https://docs.python.org/3/library/datetime.html

[attack68]

Hi @JARVIS-03
The issue wasn't really about how to to it. It was more the non-effect of a pandas options which is documented to have an effect in this regard.

[mroeschke]

I only see this explicitly called here

pandas/pandas/_libs/tslibs/parsing.pyx

Lines 315 to 319 in 4e1f51c

	if dayfirst is None or yearfirst is None:
	if dayfirst is None:
	dayfirst = get_option("display.date_dayfirst")
	if yearfirst is None:
	yearfirst = get_option("display.date_yearfirst")

So possibly the docs are incorrect. It's also concerning we don't have tests for these configurations.

Since it appears this is only for parsing, I think it's okay that both default to False