I have five data frames. I want to subtract 1.00 from all dataframes containing values equal to 1 or less than 1.
DT D1 D2
0 1.0 1.0
RE E1 E2 E3
0 1.0 1.0 0.8
FE F1 F2
0 1.0 0.63
SE S1 S2
0 1.0 1.0
DT D1 D2
RE E1 E2 E3 E1 E2 E3
0 1.0 1.0 1.0 1.0 0.45 1.0
DT D1 D2
RE E1 E2 E3 E1 E2 E3
FE F1 F2 F1 F2 F1 F2 F1 F2 F1 F2 F1 F2
0 1.0 0.0 0.0 1.0 1.0 1.0 1.0 1.0 1.0 0.0 1.0 1.0
I tried following for all dataframes, however, in some cases I get large negative values like 1.110223e-16, -2.220446e-16. How to avoid these values?
df = 1.00 - df
1.110223e-16 isn't a large, negative value. It's actually an extremely small positive value. The e-16 represents a tiny exponent.
1e-2
# 0.01
1e-4
# 0.0001
print('{:.16f}'.format(1e-16))
# 0.0000000000000001
My guess is you're running into floating point inconsistencies. For example, the 1s in your data may actually have tiny floating point fractions:
1 + 1.110223e-16
# 1.0
It's just that you don't get to see it when you print out your DataFrames.
One suggestion I would have is to use np.isclose and set these near 0 values to zero.
df.values[np.isclose(df.values, 0)] = 0