Inconsistent behavior of any(df == value) on pandas dataframe

You need to use pandas built in any instead of any from base Python:

df1.eq(1).any().any()
# True

df2.eq(1).any().any()
# True

When using any from python, it treats the data frame as an iterable/dictionary and thus only check the column names, without looking at the values of the data frame; If you simply loop through df1 and df2, you can see it only returns the column names, which is how a dictionary behaves; Since df1 contains column names of 0 and 1, any([0,1]) will return True; df2, on the other hand, contains only one column of [0], any([0]) returns False. So any(df == 1) is somewhat equivalent to any(df) or any(df.columns):

[x for x in df1]
# [0, 1]

[x for x in df2]
# [0]

In pandas better use DataFrame.any.

Numpy solutions:

print ((df1 == 1).values.any())
True
print ((df2 == 1).values.any())
True

You need to use (df2 == 1).any() instead