Python float type causing invalid result?

Asked Apr 12 '21 at 15:06

Active Apr 12 '21 at 15:10

Viewed 21 times

Imagine you have the following df:

dffinal =  {'BTW': '2.81', 'ExclBTW': '31.17', Totaal: '33.98'}

    BTW     ExclBTW Totaal  
0   2.81    31.17   33.98

dtype:

I want to do a reconciliation check on these numbers, such that Totaal equals ExclBTW + BTW:

dffinal['VATcheck'] = np.where(dffinal['Totaal'].astype('float') == (dffinal['ExclBTW'].astype(float) + dffinal['BTW'].astype(float)), True, False)

However, the VATcheck returns False, while 2.81 + 31.17 = 33.98:

     BTW    ExclBTW Totaal  VATcheck
0   2.81    31.17   33.98   False

What am I doing wrong here?

Please help!

edited Apr 12 '21 at 15:10

asked Apr 12 '21 at 15:06

Max

floating point precision. Use `np.isclose` instead of `==` comparison. – Quang Hoang Apr 12 '21 at 15:07
Why are you calling `round()` on one side of the equation but not the other? – John Gordon Apr 12 '21 at 15:08
@JohnGordon sorry that was an error – Max Apr 12 '21 at 15:10
@QuangHoang how would you implement this? – Max Apr 12 '21 at 15:10
1

`dffinal['VATcheck'] = np.isclose(dffinal['Totaal'], dffinal['ExclBTW'] + dffinal['BTW'])` given they are all floats. – Quang Hoang Apr 12 '21 at 15:11

0 Answers0