熊猫-比较正/负值

Question

I have a dataframe "df":

    x  y
0   1  -1
1  -2  -3
2   3   4
3   4   5
4   9   6

I am trying to determine what percentage of x and y values are in agreement in terms of being positive or negative. So if x is positive and y is positive, that would be a correct answer. If x and y are both negative, that would be correct. If x and y are different, then it is wrong. Is there a fast way to do this? Ultimately I just want to know what percentage of all rows have a correct answer.

(ps there are 1M+ rows in the actual dataframe)

Thank You

Answer 1

If we compare a product of x*y >= 0 - this should give us "good" rows:

In [19]: df['x'].mul(df['y']).ge(0)
Out[19]:
0    False
1     True
2     True
3     True
4     True
dtype: bool


In [20]: df.loc[df['x'].mul(df['y']).ge(0)]
Out[20]:
   x  y
1 -2 -3
2  3  4
3  4  5
4  9  6

In [21]: len(df.loc[df['x'].mul(df['y']).ge(0)])/len(df)
Out[21]: 0.8

or as proposed by @NickilMaveli a faster and more "Pandaic" version:

In [23]: df['x'].mul(df['y']).ge(0).mean()
Out[23]: 0.80000000000000004

the same idea, but this time using df.eval() method:

In [27]: df.eval('x * y >= 0').mean()
Out[27]: 0.80000000000000004