使用Pandas to_csv导出CSV时如何去除整数后的.0？

I totally get this frustration—nothing’s more annoying than seeing clean integers turn into 1520.0 in your CSV output, especially when you’ve got tons of columns to handle. Let’s walk through a few simple, scalable solutions to fix this without manually tweaking every column:

1. Quick global fix with float_format (for all float-as-integer columns) #

If every float column in your DataFrame is actually an integer (no real decimal values needed), use the float_format parameter in to_csv to force all floats to export as whole numbers:

todo.to_csv('D:\prueba.csv', sep=',', header=False, index=False, float_format='%.0f')

This tells Pandas to format all float values with zero decimal places, so 1520.0 becomes 1520 in the output. Just note: this affects every float column, so skip this if you have columns that need to keep decimals.

2. Target specific columns with formatters (for mixed column types) #

If only certain columns need the integer treatment, use the formatters parameter. For large numbers of columns, you can even generate the formatter dictionary dynamically:

Explicit column selection: #

# List columns that should be integers
int_columns = ['user_id', 'order_number', 'product_count']
formatter_dict = {col: '{:.0f}'.format for col in int_columns}

todo.to_csv('D:\prueba.csv', sep=',', header=False, index=False, formatters=formatter_dict)

Auto-detect whole-number float columns: #

If you want to automatically pick float columns that only contain whole numbers:

# Identify float columns where all values are whole numbers
whole_num_float_cols = [
    col for col in todo.select_dtypes(include='float').columns 
    if (todo[col] % 1 == 0).all()
]

formatter_dict = {col: '{:.0f}'.format for col in whole_num_float_cols}

todo.to_csv('D:\prueba.csv', sep=',', header=False, index=False, formatters=formatter_dict)

3. Cleanest long-term fix: Use nullable integer types #

If your columns have missing values (NaN) that forced them to be float type in the first place, switch to Pandas’ nullable integer type (Int64 with a capital I) to preserve integer formatting while supporting NaN:

# Convert columns to the most appropriate data types (including nullable integers)
todo = todo.convert_dtypes()

# Export without extra formatting—integers stay as integers!
todo.to_csv('D:\prueba.csv', sep=',', header=False, index=False)

The convert_dtypes() method automatically adjusts columns to their optimal type, so your CSV will output 1520 instead of 1520.0 even if there were missing values in the column.

内容的提问来源于stack exchange，提问作者Elvis Velasque