Recommended Free Tools
To change one column to a compatible, known type, assign the result of astype() back to that column: df['age'] = df['age'].astype('int64'). If the column contains text that must be interpreted as numbers, dates, or durations, use pd.to_numeric(), pd.to_datetime(), or pd.to_timedelta() instead.
Choose the right conversion method
The right method depends on whether the values already have the form you want or need to be parsed. Pandas documents DataFrame.astype() as a way to cast an object to a specified dtype: pandas.DataFrame.astype.
As an Amazon Associate I earn from qualifying purchases.
| Situation | Use | Example |
|---|---|---|
| Values already conform to a known dtype | astype() |
df['age'] = df['age'].astype('int64') |
| Text needs to be interpreted as numeric values | pd.to_numeric() |
df['amount'] = pd.to_numeric(df['amount']) |
| Text represents dates or durations | pd.to_datetime() or pd.to_timedelta() |
df['date'] = pd.to_datetime(df['date']) |
| Several columns need nullable-type inference | convert_dtypes() |
df = df.convert_dtypes() |
| The intended type is known before loading a CSV | read_csv(dtype=...) |
pd.read_csv('data.csv', dtype={'Value': float}) |
Cast a column with astype()
Use astype() when the existing values can be represented as the target type. Assign the returned Series back to the column:
Free tools Windows power users keep installed
One-click scans. No signup required.
df['age'] = df['age'].astype('int64')For multiple columns, pass a mapping of column names to dtypes. The following also demonstrates pandas’ nullable string dtype:
#1 Best Overall
df = df.astype({'age': 'int64', 'name': 'string'})The default behavior is to raise an error if a value cannot be converted. errors='ignore' instead returns the original object when conversion fails; it does not repair the problematic values. In pandas 3.0, the copy parameter is ignored and deprecated because the method uses lazy-copy behavior under Copy-on-Write. See the astype API documentation for version-specific details.
Convert numeric text and handle invalid values
For a column of strings such as "12.50", use pd.to_numeric() to parse the values:
df['amount'] = pd.to_numeric(df['amount'])By default, invalid numeric text raises an error. If you choose errors='coerce', unparseable entries become missing values. Inspect those results so malformed data is not silently overlooked:
df['amount'] = pd.to_numeric(df['amount'], errors='coerce')
invalid_or_missing = df['amount'].isna()downcast can request a smaller suitable numeric dtype, using options such as 'integer', 'signed', 'unsigned', or 'float'. Validate the values and precision you need: pandas warns that very large numbers may lose precision because of ndarray representation limits. Refer to the to_numeric API.
Parse dates and durations
Use the dedicated parsing functions when strings represent dates or elapsed time:
df['date'] = pd.to_datetime(df['date'])
df['elapsed'] = pd.to_timedelta(df['elapsed'])A direct cast is not a substitute for interpreting arbitrary date text. Inconsistent or unparseable date values can also prevent a clean datetime result; validate the converted column against the input. Pandas lists these functions among its general functions.
Preserve missing values with nullable dtypes
Ordinary NumPy integer dtypes cannot represent missing values. If an integer column must retain missing entries, consider pandas’ nullable integer dtype Int64 (capital “I”), after ensuring the non-missing values are valid integers:
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchPC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11df['count'] = pd.to_numeric(df['count'], errors='raise').astype('Int64')Pandas also provides nullable string and boolean types. Nullable types use pd.NA to represent missing values. See the missing-data guide and numeric conversion documentation.
Infer nullable dtypes across a DataFrame
When you want pandas to choose nullable string, boolean, integer, and floating types rather than specify one exact dtype, use convert_dtypes():
df = df.convert_dtypes()This returns a DataFrame with attempted nullable-type conversions; it is not a guarantee that every column will become a particular type. Its dtype_backend choices include 'numpy_nullable' and 'pyarrow'. Pandas marks this backend option experimental, so consult the convert_dtypes API when choosing it.
Set a column’s type while reading a CSV
If the intended dtype is known before import, specify it with read_csv(dtype=...):
df = pd.read_csv('data.csv', dtype={'Value': float})Mixed values in a CSV can lead to an object column and a DtypeWarning. Setting dtype can make the intended interpretation explicit; converters or a post-read conversion may be more appropriate when the input needs special handling. For dates, pandas provides date-parsing support, but inconsistent or unparsable values may prevent a datetime result. See the read_csv API and pandas’ IO guide.
Check the result and choose a recovery path
After conversion, inspect the dtype and look for values that became missing or caused conversion to fail:
Quick Recap
print(df['amount'].dtype)
print(df['amount'].isna().sum())- Known values, known target: use
astype(); correct values that do not conform rather than relying on an ignored error. - Textual numbers: parse with
to_numeric(); choose whether invalid entries should raise or become missing, then inspect them. - Dates or elapsed times: use the corresponding datetime or timedelta parser and check inconsistent input.
- Missing integers or booleans must remain missing: choose nullable dtypes such as
Int64rather than ordinary NumPy integer types. - Import inference is the issue: specify a dtype or converter at read time when the intended interpretation is known.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
