Pandas Python: why does np.round() 3.601626 to 2 dp returns 3.6?

Question

I'm trying to round some floats in a df to 2 dp but all multiples of 10 are rounded to 1 dp, so 3.60 would be 3.6 and 12.40 would be 12.4...

I have a df:

Values           A         B         C         D
Question                                                                                                    
A2         4.08642  4.144279  3.601626  3.983852

my code

np.round(df,2)

output

Values        A     B    C     D
Question                                                                                                    
A2         4.09  4.14  3.6  3.98

expected output

Values        A     B     C     D
Question                                                                                                    
A2         4.09  4.14  3.60  3.98

how can I force np.round to display '3.6' as '3.60'?

FYI - I don't want to use a string formatter because I have to round these numbers.

"I don't want to use a string formatter because I have to round these numbers" - how are those two things in any way related? Why would having to round the numbers make you not want to use string formatting? — user2357112, Apr 14 '15 at 09:27

score 4 · Accepted Answer · edited May 23 '17 at 11:50

4

The problem is not with the rounding, but with the way pandas displays the DataFrame.

You can set float_format like:

pd.options.display.float_format = '${:,.2f}'.format

As explained in more details in this answer.

edited May 23 '17 at 11:50

Community

1
1

answered Apr 14 '15 at 09:18

shx2

61,779
13
130
153

This works at a superficial level as doing mean_df.values still returns [[ 4.09 4.14 3.6 3.98]] – Boosted_d16 Apr 14 '15 at 10:17

score 1 · Answer 2 · answered Apr 14 '15 at 09:25

1

You can use string formater and then use Decimal function form library decimal to return number again:

Example code:

import decimal
print(decimal.Decimal("{:.2f}".format(3.6)))

answered Apr 14 '15 at 09:25

Julia K

46
7

I don't think the OP wants to change the underlying dtype which decimal will do – EdChum Apr 14 '15 at 09:56

score 1 · Answer 3 · answered Apr 14 '15 at 10:08

1

Another way to set the pandas display.float_format option:

>>> pd.set_option('float_format', '{:,.2f}'.format)

answered Apr 14 '15 at 10:08

Vidhya G

2,250
1
25
28

2

How is this different from shx2's answer? – EdChum Apr 14 '15 at 10:11
@edchum one uses `pd.options.display.float_format = ` and the other uses `pd.set_option(...)`. – Vidhya G Apr 14 '15 at 10:14
They are both the same underlying option – EdChum Apr 14 '15 at 10:15
1

Different Syntax. Some prefer one over the other. – Vidhya G Apr 14 '15 at 10:21
Thanks, had to go with shx2 because he was first to answer. – Boosted_d16 Apr 14 '15 at 14:32
No problem. If it was in any way helpful to you, or anyone else in the future, then job done. – Vidhya G Apr 14 '15 at 14:35

Pandas Python: why does np.round() 3.601626 to 2 dp returns 3.6?

3 Answers3