Numpy diff calculates the n-th order discrete difference along the given axis. It calculates the first-order difference using the formulae * result[n] = a[n+1] – a[n]* .Here

**a**denotes NumPy array. The second or higher-order differences are calculated recursively.

In this entire article, you will know how to implement NumPy diff with various examples.

But before moving to the coding examples let’s learn the syntax of the * numpy.diff()* method.

## The syntax of numpy diff() method

`numpy.diff(a, n=1, axis=-1)`

**a:** Input array

**n:** Order of difference. The number of times values are differenced.

**axis:** The axis along which the difference is taken, default is the last axis.

The method returns the difference of array type along the axis.

## NumPy diff of 1D Array

In this section, you will learn to calculate the difference of 1 D Numpy array. Let’s create the array.

`array_1d = np.array([10,20,30,40,50,60,70])`

Just pass the array inside the * diff()* method.

`np.diff(array_1d)`

The above code will calculate the first-order difference. To calculate the 2,3, or n difference you have to pass the n =2,3,4,5, so on.

**NumPy 2nd order diff**

`np.diff(array_1d,n=2)`

### 3rd Order Difference

`np.diff(array_1d,n=3)`

**Output**

## NumPy difference of 2D Array

The same * diff()* method calculates the difference of two or N-dimensional array. Here you can also calculate numpy difference along the axis that is row-wise or column-wise. Let’s create a Two-dimension NumPy array.

`array_3x4 = np.array([[10,20,30,40],[50,60,70,80],[90,100,110,120]])`

### 1st order diff

`np.diff(array_3x4)`

The above code will find the first-order difference column-wise. If you want to do calculation row-wise then you have to pass the axis value to 0.

### NumPy diff of rows

`np.diff(array_3x4,axis=0)`

Just like you are finding the 2,3 or n order in one-dimensional array. You can also find it here bypassing the n=2,3,4 or so on.

**Output**

## Other Examples

### Numpy Difference for Time-Series Data

If You have time-series then you can also calculate the difference over it. Let’s create a time series array for date of datetime type.

`data = np.arange('2020-11-01', '2020-11-16', dtype=np.datetime64)`

Just pass the data inside the * diff() *method to get the output.

`np.diff(data)`

**Output**

## Conclusion

These are examples for finding numpy difference for both one and two-dimensional array. The application of finding a difference is that you can find some useful patterns in the dataset that helps you do predict results. For example in stock markers.

Hope you have liked this tutorial. If you have any queries then you can contact us.

#### Join our list

Subscribe to our mailing list and get interesting stuff and updates to your email inbox.