Convert Floats to Integers in a pandas DataFrame

In this tutorial, you will convert DataFrame columns of type float to integer.

TLDR solution

df['column_a'] = df['column_a'].astype('int')

# if column_a has NaN values
df['column_a'] = df['column_a'].fillna(0).astype('int')

Example 1: Convert Floats to Integers in a Specific Column

Let's say, you have the following DataFrame:

data = { "fish": ["salmon", "pufferfish", "shark", "mackerel"],
         "count": [2.0, 35, 53.0, 88]}

df = pd.DataFrame(data)

print(df)
print(df.dtypes)

         fish  count
0      salmon    2.0
1  pufferfish   35.0
2       shark   53.0
3    mackerel   88.0
fish      object
count    float64
dtype: object

That is, column count is of type float.

You can use the astype function to convert the floats to integers:

df['count'] = df['count'].astype('int')

print(df)
print(df.dtypes)

         fish  count
0      salmon      2
1  pufferfish     35
2       shark     53
3    mackerel     88
fish     object
count     int64
dtype: object

Example 2: When all Columns are of Type Float

Suppose:

data = { "size": [50.0, 22.0, 39.0, 20.0],
         "count": [2.0, 35.0, 53.0, 88.0]}

df = pd.DataFrame(data)

print(df)
print(df.dty

   size  count
0  50.0    2.0
1  22.0   35.0
2  39.0   53.0
3  20.0   88.0
size     float64
count    float64
dtype: object

Convert the entire DataFrame like this:

df = df.astype('int')

print(df)
print(df.dtypes)

   size  count
0    50      2
1    22     35
2    39     53
3    20     88
size     int64
count    int64

Example 3: Convert Floats in a Mixed Type DataFrame

Suppose:

data = { "fish": ["salmon", "pufferfish", "shark", "mackerel"],
         "size": [50.0, 22.0, 39.0, 20.0],
         "count": [2.0, 35.0, 53.0, 88.0]}

df = pd.DataFrame(data)

print(df)
print(df.dtypes)

         fish  size  count
0      salmon  50.0    2.0
1  pufferfish  22.0   35.0
2       shark  39.0   53.0
3    mackerel  20.0   88.0
fish      object
size     float64
count    float64
dtype: object

Convert the two columns of type float like this:

df[['size', 'count']] = df[['size', 'count']].astype('int')

print(df)
print(df.dtypes)

         fish  size  count
0      salmon    50      2
1  pufferfish    22     35
2       shark    39     53
3    mackerel    20     88
fish     object
size      int64
count     int64
dtype: object

Example 4: Convert Floats to Integers When NaN Values are present

Suppose:

data = { "fish": ["salmon", "pufferfish", "shark", "mackerel"],
         "count": [2.0, 35, np.nan, 88]}

df = pd.DataFrame(data)

print(df)
print(df.dtypes)

         fish  count
0      salmon    2.0
1  pufferfish   35.0
2       shark    NaN
3    mackerel   88.0
fish      object
count    float64
dtype: object

Replace NaN values with zeros first using fillna(0), then convert:

df['count'] = df['count'].fillna(0).astype('int')

print(df)
print(df.dtypes)

         fish  count
0      salmon      2
1  pufferfish     35
2       shark      0
3    mackerel     88
fish     object
count     int64
dtype: object

That's it! You just converted floats to integers in a DataFrame.