type switch from int to float64 after merge creating error

Question

I'm trying to merge two dataframes in Pandas. One of the dataframes has a numerical column whose type is "int64"

However, after the merge, the type is switched to "float64" for some reason. Note that this is not my join column

When I try to access the dataframe, it errors out:

In [56]: account_aggregates.head()
Out[56]: ) failed: TypeError: %d format: a number is required, not numpy.float64>

Answer 1

The reason that the dtype is changed to float64 is because missing values NaN cannot be represented using integer.

With respect to the error message, I had a hunch that it was 'display.float_format' as I answered a question earlier today on this and saw this error. I think that it's because you have to pass a str.format as the value rather than a format string:

pd.set_option('display.float_format', '{:.2g}'.format) as opposed to pd.set_option('display.float_format', '%.2g') as an example.