The **`to_numeric` function** in pandas is used to convert the given argument to a numeric type. The argument is converted to `float64` or `int64` by default, based on what is provided in the arguments. We can use the `downcast` parameter if we want to convert data to a particular type.

The illustration below shows how the `to_numeric` function works:

# Syntax

The syntax of the `to_numeric` function is as follows:

```python
pd.to_numeric(arg, errors, downcast)
```
# Parameters

The `to_numeric` function has the following parameters:

| Parameter  | Description  |
| - | - |
| `arg`  |  Argument to be converted. Can be scalar, list, tuple, 1-d array, or Series.  |
| `errors`  | Determines how to handle errors during conversions. Can be `ignore`, `raise` or `coerce`. It is `raise` by default.  |
| `downcast`  | Specifies conversion to a particular datatype. Can be `int`, `signed`, `unsigned` or `float`. By default, it is `None`. If it is not `None`, pandas will downcast the data to the smallest data type possible. |

> Only the `arg` parameter is compulsory. The rest are optional.

# Return value

Data in the converted format is returns based on what is passed in the argument.

# Handling errors

Different types of error handling are possible within the `to_numeric` function:

* If the error argument is passed as `raise`, then invalid parsing will raise an exception.

* If the error argument is passed as `coerce`, then invalid parsing will be set as `NaN`.

* If the error argument is passed as `ignore`, then invalid parsing will return the input.

# Example

The code snippet below shows how we can use the `to_numeric` function in pandas:

import pandas as pd
import numpy as np

s = pd.Series(['1.0', '2', -3])
print("Series in string")
print(s)
print("Default Numeric Form")
print(pd.to_numeric(s))
print('\n')

print("Numeric Form using downcast = signed")
print(pd.to_numeric(s, downcast='signed'))
print('\n')

print("Numeric Form using downcast = int")
print(pd.to_numeric(s, downcast='integer'))
print('\n')

print("Ignoring errors")
s2 = pd.Series(['apple', '1.0', '2', -3])
print(pd.to_numeric(s2, errors='ignore'))
print('\n')

print("Corerce errors")
s2 = pd.Series(['apple', '1.0', '2', -3])
print(pd.to_numeric(s2, errors='coerce'))
print('\n')


Parameter	Description
`arg`	Argument to be converted. Can be scalar, list, tuple, 1-d array, or Series.
`errors`	Determines how to handle errors during conversions. Can be `ignore`, `raise` or `coerce`. It is `raise` by default.
`downcast`	Specifies conversion to a particular datatype. Can be `int`, `signed`, `unsigned` or `float`. By default, it is `None`. If it is not `None`, pandas will downcast the data to the smallest data type possible.

What is the to_numeric function in pandas?

Syntax

Parameters

Return value

Handling errors

Example