TypeError while converting pandas column to int

Question

Working with pandas data frame, where one of the columns, say col1 has floating point values, I am trying to divide each of these values with a pre defined constant, say A, and then save the results as integer values.

A = 0.5

Following is the data in col1

df["col1"]

0     0.800000
1     0.883333
2     0.883333
3     1.000000
4     1.000000
5     1.300000
6     1.300000
7     1.500000
8     1.500000
9     2.000000
10    2.000000
11    2.500000
12    2.500000

After applying

df["new_col"] = (df["col1"] / A)

It gives

0     1.600000
1     1.766667
2     1.766667
3     2.000000
4     2.000000
5     2.600000
6     2.600000
7     3.000000
8     3.000000
9     4.000000
10    4.000000
11    5.000000
12    5.000000

which is fine, but as soon as I add .astype(int) to the above code, it can be observed at index 9 and 10 the values are 3 and 3 whereas it should be 4 and 4 respectively.

df["new_col"] = (df["col1"] / A).astype(int)
df["new_col"]

0     1
1     1
2     1
3     2
4     2
5     2
6     2
7     3
8     3
9     3
10    3
11    5
12    5

The other ways I am trying to obtain the result are

 df["new_col"] = math.floor(df["col1"] / A )

and

df["new_col"] = int( df["col1"] / A)

Both of which gives me TypeError saying:

TypeError: cannot convert the series to class 'float' and TypeError: cannot convert the series to class 'int' respectively.

Please let me know how should I resolve the above issues.

I guess, I should restart my system then and try again, otherwise I've tried the code 10 times and it shows me 3 and 3. — Ankie
– Ankie, Commented Aug 9, 2017 at 21:30
That's probably a floating point issue. If it's represented as 3.99999 astype will round it down to 3. Try round(0) maybe? — user2285236
– user2285236, Commented Aug 9, 2017 at 21:32
@ayhan I did what u suggested, using .round(0) , but the problem is it works as a ceiling function, it makes values like 2.6, 2.87 to 3, whereas I want them to be 2. What else can I try ? — Ankie
– Ankie, Commented Aug 9, 2017 at 22:01

B. M. · Accepted Answer · 2017-08-09 21:35:02Z

0

You probably have rounding issue. What you see as 4.000000 is probably 3.9999999999 internally, so you get this result. ( try df.col1-2 to check it.)

You can try (df.round(6)/.5).astype(int) to work on the digits you see, but it is a workaround.

answered Aug 9, 2017 at 21:35

B. M.

18.7k2 gold badges40 silver badges56 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Ankie Over a year ago

What if all the values in the above column are negative, so in case of -3.999999, it will be rounded to -4 ?

B. M. Over a year ago

It's dangerous to continue this way. I think if your computations lead to 3.999 when 4 is the good result, you must refactor this part.

Ankie Over a year ago

I am still struggling with this, now one of the values in column is -7.680000 and I am dividing it with 0.04000 , it should be -192, but I get -191 which affects my overall result. Please suggest the refactoring part yo mentioned. I am sorry, haven't encountered this type of problem before, so just running out of solutions. I have a column which has only negative floating point numbers which I will be dividing with another positive floating point no, and then keeping only the integer part of the result (like floor and not ceiling).

Collectives™ on Stack Overflow

TypeError while converting pandas column to int

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related