version 0.11 issues with to_csv

[superkeyor]

It looks like that some changes were made to the DataFrame.to_csv method, as mentioned in the what's new document.

Today I found two issues with the updated to_csv:

1. I have an int column, when I use floatformat parameter, the int are changed to float
2. cols parameter simply changes the name of columns, it does not reorder the data in a whole column.

Thanks for your hard working. Hope we can make pandas less buggy and more powerful!

[ghost]

2 was fixed in master a few days ago and will be in 0.11.1 #3454, the bugfix release due in a few weeks.

[jreback]

can you give an example of your 1)?

[jreback]

@jerryzhujian9 can you give us an example for 1)?

[ghost]

This works just fine:

df.to_csv("/tmp/1.csv",float_format="%.2f")
pd.read_csv("/tmp/1.csv").info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 10 entries, 0 to 9
Data columns (total 3 columns):
R0         10  non-null values
C_l0_g0    10  non-null values
C_l0_g1    10  non-null values
dtypes: int64(2), object(1)

Closing as can't repro. @jerryzhujian9 will gladly reopen if you provide steps to reproduce.