Author name: Satyabrata Jena

In this tutorial, python beginners and experts can easily learn about the difference between numpy ravel and flatten functions. The main aim of both numpy.ravel() and numpy.flatten() functions are the same i.e. to flatten a numpy array of any shape. Before learning about flatten ravel differences, let’s take a look at the basic overview of both the numpy functions from here.

numpy.ravel()

ravel() is a built-in function in the module of numpy that returns a view of the original array and also returns a flattened 1D view of the numpy array object. Moreover, it accepts an array-like element as a parameter.

Syntax of the ravel() function in numpy:

numpy.ravel(a, order='C')

numpy.ndarray.flatten()

Flatten always returns a copy. The function of the flatten() is a faster member function of the numpy array where it returns a flattened 1D copy of the numpy array object.

Syntax of Numpy flatten() function:

ndarray.flatten(order='C')

Let’s discuss some of the differences between the numpy ravel & flatten function:

Differences between ravel() & flatten()

First of all, import the numpy module,

import numpy as np

A quick view about the difference between flatten and ravel functions in numpy is furnished below:

a.flatten()

(i) Return copy of the original array

(ii) If you modify any value of this array value of the original array is not affected.

(iii) Flatten() is comparatively slower than ravel() as it occupies memory.

(iv) Flatten is a method of a ndarray object.

a.ravel()

(i) Return only reference/view of original array

(ii) If you modify the array you would notice that the value of the original array also changes.

(iii) Ravel is faster than flatten() as it does not occupy any memory.

(iv) Ravel is a library-level function.

For in-depth knowledge on the difference between flatten and ravel numpy functions, kindly check out the further sections without any fail.

Difference 1: Performance: Copy vs view

Thenumpy.ndarray.flatten()function returns a flatten copy of numpy array object only, whereas numpy.ravel() returns a flattened view of the input array if possible.

import numpy as sc
# Creation of 2D Numpy array
arr_twoD = sc.array([[10, 20, 30],
                [40, 50, 60],
                [70, 85, 90]])
print('2D array')                
print(arr_twoD)
# Convert the 2D array to flattened 1d view
flatn_arr = arr_twoD.flatten()
print('Flattened 1D view')
flatn_arr[1] = 55
print(flatn_arr)
print(arr_twoD)

Output :
2D array
[[10 20 30]
 [40 50 60]
 [70 85 90]]
Flattened 1D view
[10 55 30 40 50 60 70 85 90]
[[10 20 30]
 [40 50 60]
 [70 85 90]]

Here we can successfully change 2nd element of the numpy array that is reflected in the 1D numpy array. But the original 2D array is unaffected which denotes that flatten() returns a copy of the input array.

Let’s try using numpy.ravel() to convert 2D numpy array to 1D array

ravel() returns a flattened view of the numpy array object if possible. When there is any change done in the view object it will also get reflected in the original numpy array.

import numpy as sc
# Creation of 2D Numpy array
arr_twoD = sc.array([[10, 20, 30],
                [40, 50, 60],
                [70, 85, 90]])
print('2D array')                
print(arr_twoD)
# Convert the 2D array to flattened 1d view
flatn_arr = sc.ravel(arr_twoD)
flatn_arr[1] = 55
print('Flattened 1D view:')
print(flatn_arr)
print(arr_twoD)

Output :
2D array
[[10 20 30]
 [40 50 60]
 [70 85 90]]
Flattened 1D view:
[10 55 30 40 50 60 70 85 90]
[[10 55 30]
 [40 50 60]
 [70 85 90]]

We can check of the ravel()function returns a view object or not by using the base attribute of the returned object. If it is None then it returns to the original numpy array i.e. it returns a flattened array which is a view only.

import numpy as sc
# Creation of 2D Numpy array
arr_twoD = sc.array([[10, 20, 30],
                [40, 50, 60],
                [70, 85, 90]])
print('Original 2D array')                
print(arr_twoD)
# Convert the 2D array to flattened 1d view
flatn_arr = sc.ravel(arr_twoD)
if flatn_arr.base is not None:
    # Change the 2nd element of flat array
    flatn_arr[1] = 55
    # Modification will be reflected in 2D numpy array and flattened array
    print('Flattened 1D view:')
    print(flatn_arr)
    print(arr_twoD)

Output :
Original 2D array
[[10 20 30]
 [40 50 60]
 [70 85 90]]
Flattened 1D view:
[10 55 30 40 50 60 70 85 90]
[[10 55 30]
 [40 50 60]
 [70 85 90]]

So ravel() returns a view of the input array and hence here the performance of ravel() is better than flatten().

Difference 2: Compatibility with other array-like sequences (list etc)

The ndarray.flatten()function can be used to flatten a numpy array object only. In the case of a numpy.ravel()function being a built-in function, it accepts an array-like element, so we can even pass a list to it.

import numpy as sc
# Creation of list of lists
list_lists = [[10, 20,30],
[40, 50, 60],
[70, 85, 90]]
# Creation of flattened numpy array from list of lists
flatn_arr = sc.ravel(list_lists)
print('Flattened 1D Array:')
print(flatn_arr)

Output :
Flattened 1D Array:
[10 20 30 40 50 60 70 85 90]

While this is not feasible with ndarray.flatten() function.

Conclusion on A view of the original array using numpy ravel

Therefore, to conclude in the end there are 2 main differences between the ndarray.flatten() and numpy.ravel() function,

1. ravel() function returns a view of the array if possible, where flatten() always returns a copy. Therefore the performance of ravel() is much better than flatten()
2. ravel() function can accept other array-like elements like lists etc. Although, flatten() can work with numpy arrays only.

In this tutorial, we will discuss about how we can remove items/rows/columns from the Numpy array. We can remove elements from a NumPy array by index position using numpy.delete( ). Get to know the syntax of numpy.delete() function and learn about all its parameters with a clear explanation. Refer to the further modules to be aware of procedures on how to remove or delete an element or multiple elements in rows/columns of 1D and 2D Arrays by its Index Position.

• How to remove items/rows/columns from Numpy array using np.delete() in Python?
• Delete an Element in 1D Numpy Array by its Index position
• Delete multiple elements in 1D Numpy Array by its Index position
• Deleting rows & columns from a 2D Numpy Array
• Delete a column in 2D Numpy Array by its column number
• Delete multiple columns in 2D Numpy Array by its column number
• Delete a row in 2D Numpy Array by its row number
• Delete multiple rows in 2D Numpy Array by its row number
• Delete specific elements in 2D Numpy Array by its index position

How to remove items/rows/columns from Numpy array using np.delete() in Python?

Python provides a method to delete elements from the numpy array depending on the index position and its syntax is provided below.

numpy.delete( )

Python’s NumPy library has a method to delete elements. It is done by using the np.delete( ) function.

Syntax:

numpy.delete(arr, obj, axis=None)

Where,

• arr: Array from which elements are to be deleted
• obj: Index position or positions from which elements are to be deleted
• axis: The axis along which we want to delete (1 means delete along columns, 0 means delete along the rows, if None then the array is flattened out and then the elements are to be deleted)

It returns a copy of the array with the required elements deleted.

Delete an Element in 1D Numpy Array by its Index position

Let’s see the implementation of it.

#Program :

import numpy as np
#Numpy array
arr = np.array([10,20,30,40,50,60,70,80,90])
#deleting a=item at index 2
arr = np.delete(arr, 2)
print('Modified Numpy Array after deleting element at index 2')
print(arr)

Output :
Modified Numpy Array after deleting element at index 2
[10 20 40 50 60 70 80 90]

Delete multiple elements in 1D Numpy Array by its Index position

Let’s see the implementation of it.

#Program : 

import numpy as np 
#Numpy array 
arr = np.array([10,20,30,40,50,60,70,80,90]) 
#deleting a=item at index 2,5,7 
arr = np.delete(arr, 2) 
print('Modified Numpy Array after deleting element at index 2,5,7') 
print(arr)

Output :
Modified Numpy Array after deleting element at index  2,5,7
[10 20 40 50 70 90]

Deleting rows & columns from a 2D Numpy Array

Delete a column in 2D Numpy Array by its column number

if we want to delete a column from a 2D numpy array using np.delete() then we have to pass the axis=1 along with numpy array and index of column.

Let’s see the implementation of it.

#program

import numpy as np

arr2d = np.array([[10,20,30],
[40,50,60],
[70,80,90]])
#deleting elements at column 1
arr = np.delete(arr2d, 1, axis=1)
print('Modified Numpy Array by deleting element at index position 2,5,7')
print(arr)

Output :
Modified Numpy Array by deleting element at index position 2,5,7
[[10 30]
[40 60]
[70 90]]

Delete multiple columns in 2D Numpy Array by its column number

To delete multiple columns pass axis=1 and list of column numbers to be deleted along with numpy array to the function.

Let’s see the implementation of it.

#Program :

import numpy as np

arr2D = np.array([[11 ,12, 13, 11],
                [21, 22, 23, 24],
                [31, 32, 33, 34]])
arr2D = np.delete(arr2D, [2,3], axis=1)
print('Modified 2D Numpy Array by removing columns at index 2 and 3')
print(arr2D)

Output :
Modified 2D Numpy Array by removing columns at index 2 and 3
[[11 12]
[21 22]
[31 32]]

Delete a row in 2D Numpy Array by its row number

To delete a row from a 2D numpy array using np.delete() we need to pass the axis=0 along with numpy array and the row index.

Let’s see the implementation of it.

#Program : 

import numpy as np 
arr2D = np.array([[11 ,12, 13, 11], [21, 22, 23, 24], [31, 32, 33, 34]]) 
arr2D = np.delete(arr2D, 0, axis=0) 
print('Modified 2D Numpy Array by removing rowss at index 0') 
print(arr2D)

Output :
Modified 2D Numpy Array by removing rowss at index 0
[[21 22 23 24]
[31 32 33 34]]

Delete multiple rows in 2D Numpy Array by its row number

To delete multiple rows pass axis=0 and list of row numbers to be deleted along with numpy array to np.delete()

#Program :

import numpy as np

arr2D = np.array([[11 ,12, 13, 11],
                [21, 22, 23, 24],
                [31, 32, 33, 34]])
arr2D = np.delete(arr2D, [1, 2], axis=0)
print('Modified 2D Numpy Array by removing rows at index 1 and 2')
print(arr2D)

Output :
Modified 2D Numpy Array by removing rows at index 1 and 2
[[11 12 13 11]]

Delete specific elements in 2D Numpy Array by its index position

When we don’t mention axis value, the default value is none which means the array gets flattened. After that, we use np.delete() to delete elements from rows and columns. The function will return a flattened array without the deleted rows and column elements.

#program :

import numpy as np

arr2D = np.array([[11 ,12, 13, 11],
                [21, 22, 23, 24],
                [31, 32, 33, 34]])
arr2D = np.delete(arr2D, 2)
print('Modified 2D Numpy Array by removing element at row 0 column 2')
print(arr2D)

Output :
Modified 2D Numpy Array by removing element at row 0 column 2
[11 12 11 21 22 23 24 31 32 33 34]

NumPy arrays are very essential data structures for working with data in Python, machine learning models. Python’s Numpy module provides a function to save a numpy array to a txt file with custom delimiters and other custom options. In this tutorial, we will discuss the procedure of how to save Numpy Array to a CSV File with clear steps.

• numpy.savetxt() Function
• How to save Numpy Array to a CSV File using numpy.savetxt() in Python?
• Example Program for Saving a Numpy Array to a CSV File
• Save 1D Numpy array to CSV file with Header and Footer

numpy.savetxt() Function

The numpy.savetxt() function is the counterpart of the NumPy loadtxt() function and can save arrays in delimited file formats such as CSV. Save the array we created with the following function call:

Synatx : numpy.savetxt(fname, array_name, fmt='%.18e', delimiter=' ', newline='\n', header='', footer='', comments='# ', encoding=None)

Where,

• fname: If the filename ends in .gz, the file is automatically saved in compressed gzip format. The loadtxt() function can understand gzipped files transparently.
• arr_name: It indicates data to be saved like 1D or 2D numpy array.
• fmt: It refers to a formatting pattern or sequence of patterns, which will be used while saving elements to file.
• delimiter: It is optional, refers to string or character to be used as element separator
• newline: It is optional, refers to string or character to be used as line separator
• header: It refers to a string that is written at the beginning of the file.
• footer: It refers to a string that to be written at the end of the txt file.
• comments: It refers to a custom comment marker, where the default is ‘#’. It will be pre-appended to the header and footer.

How to save Numpy Array to a CSV File using numpy.savetxt() in Python?

One of the most common file formats for storing numerical data in files is the comma-separated variable format or CSV in Short. Usually, input data are stored in CSV Files as it is one of the most convenient ways for storing data.

Savetxt function is used to save Numpy Arrays as CSV Files. The function needs a filename and array as arguments to save an array to CSV File. In addition, you need to mention the delimiter; for separating each variable in the file or most commonly comma. You can set via the “delimiter” argument.

Example Program on How to Save a Numpy Array to a CSV File

#Program :

import numpy as np
def main():
   # Numpy array created with a list of numbers
   array1D = np.array([9, 1, 23, 4, 54, 7, 8, 2, 11, 34, 42, 3])
   print('Real Array : ', array1D)
   print('<** Saved 1D Numpy array to csv file **>')
   # Save Numpy array to csv
   np.savetxt('array.csv', [array1D], delimiter=',', fmt='%d')
   print('*** Saving 1D Numpy array to csv file with Header and Footer ***')
   # Saving Numpy array to csv with custom header and footer
   np.savetxt('array_hf.csv', [array1D], delimiter=',', fmt='%d' , header='A Sample 2D Numpy Array :: Header', footer='This is footer')

   print('*** Saving 2D Numpy array to csv file ***')
   # A 2D Numpy array list of list created
   array2D = np.array([[111, 11, 45, 22], [121, 22, 34, 14], [131, 33, 23, 7]])
   print('2D Numpy Array')
   print(array2D)
   # Saving 2D numpy array to csv file
   np.savetxt('2darray.csv', array2D, delimiter=',', fmt='%d')
   # Saving 2nd column of 2D numpy array to csv file
   np.savetxt('2darray_column.csv', [array2D[:,1]], delimiter=',', fmt='%d')
   # Saving 2nd row of 2D numpy array to csv file
   np.savetxt('2darray_row.csv', [array2D[1] ], delimiter=',', fmt='%d')

   # Creating the type of a structure
   dtype = [('Name', (np.str_, 10)), ('Marks', np.float64), ('GradeLevel', np.int32)]
   #Creating a Strucured Numpy array
   structuredArr = np.array([('Sam', 33.3, 3), ('Mike', 44.4, 5), ('Aadi', 66.6, 6), ('Riti', 88.8, 7)], dtype=dtype)
   print(structuredArr)
   # Saving 2D numpy array to csv file
   np.savetxt('struct_array.csv', structuredArr, delimiter=',', fmt=['%s' , '%f', '%d'], header='Name,Marks,Age', comments='')
if __name__ == '__main__':
  main()

Output:

Real Array :  [ 9  1 23  4 54  7  8  2 11 34 42  3]
<** Saved 1D Numpy array to csv file **>
<** Saved 1D Numpy array to csv file with custom  Header and Footer **>
<** Save 2D Numpy array to csv file **>
* 2D Numpy Array *
[[111  11  45  22]
[121  22  34  14]
[131  33  23   7]]
[('Rags', 33.3, 3) ('Punit', 44.4, 5) ('Drishti', 66.6, 6)  ('Ritu', 88.8, 7)]

The Passed Delimeter ‘,’ will change to CSV Format. In addition, the format string %d passed will store the elements as integers. By default, it will store numbers in float format. Keep in mind that if you don’t mention [] around numpy array to change it to list while passing numpy.savetxt() comma delimiter willn’t work and uses ‘\n’ by default. Thus, surrounding array by [] i.e. [arr] is mandatory.

Save 1D Numpy array to CSV file with Header and Footer

In order to add comments to the header and footer while saving to a CSV File, we can pass the Header and Footer Parameters as such

# Save Numpy array to csv with custom header and footer
np.savetxt('array_hf.csv', [arr], delimiter=',', fmt='%d' , header='A Sample 2D Numpy Array :: Header', footer='This is footer')

Usually, By default comments in both the header and footer are pre-appended by ‘#’. To change this we can pass the parameter comments like comments=’@’.

Final Words

Numpy savetxt can be an extremely helpful method for saving an array to CSV File. If you want to manipulate or change the existing data set this can be a great method. If you have any queries don’t hesitate to ask us via comment box so that we can get back to you at the soonest possible. Bookmark our site for the latest updates on Python, Java, C++, and Other Programming Languages.

Do you want to determine whether a process is running or not by name and find its Process ID as well? Then, this is the right place as this article will shed light on you completely on how to check if a process is running or not using Name. We have provided enough examples for finding the Process ID PID by Name for better understanding. You can visit our site online to seek guidance or simply copy-paste the code from our portal and try to apply this knowledge and write a code on your own.

• How to Check if there exists a Process running by Name and find its Process ID (PID)?
• Determine if Process is Running or Not
• Finding the Process ID(PID) by Name?

How to Check if there exists a Process running by Name and find its Process ID (PID)?

In this article, we will discuss finding a running process PIDs by name using psutil.

psutil library can be installed by using

pip install psutil

Determine if Process is Running or Not

Follow the simple guidelines for checking if a process is running or not. They are as such

• def findProcessIdByName(processName):
• for proc in psutil. process_iter():
• pinfo = proc. as_dict(attrs=[‘pid’, ‘name’, ‘create_time’])
• if processName. lower() in pinfo[‘name’]. lower() :
• except (psutil.NoSuchProcess, psutil.AccessDenied , psutil.ZombieProcess) :

To check if a process is running or not we will iterate over all the running processes using psutil.process_iter() and during iteration match the process name.

So let’s see this example to check chrome process is running or not.

import psutil
def checkProcessRunning(processName):
    # Checking if there is any running process that contains the given name processName.
    #Iterate over the all the running process
    for proc in psutil.process_iter():
        try:
            # Check if process name contains the given name string.
            if processName.lower() in proc.name().lower():
                return True
        except (psutil.NoSuchProcess, psutil.AccessDenied, psutil.ZombieProcess):
            pass
    return False;

# Checking if any chrome process is running or not.
if checkProcessRunning('chrome'):
    print('Yes chrome process was running')
else:
    print('No chrome process was running')

Output:

No chrome process was running

As in my system chrome instances are not running it will return False.

Finding the Process ID(PID) by Name?

In the above program, we checked any chrome instances were running or not. But in this example, we will see how to get the Process ID of a running ‘chrome.exe’ process. For that we will iterate over all the running processes and during iteration for each process whose name contains the given string, we will keep its information in a list. The following code helps you find the PID of a running process by Name.

import psutil
def findProcessIdByName(processName):
    # Here is the list of all the PIDs of a all the running process 
    # whose name contains the given string processName
    listOfProcessObjects = []
    #Iterating over the all the running process
    for proc in psutil.process_iter():
       try:
           pinfo = proc.as_dict(attrs=['pid', 'name', 'create_time'])
           # Checking if process name contains the given name string.
           if processName.lower() in pinfo['name'].lower() :
               listOfProcessObjects.append(pinfo)
       except (psutil.NoSuchProcess, psutil.AccessDenied , psutil.ZombieProcess) :
           pass
    return listOfProcessObjects;

# Finding PIDs od all the running instances of process 
# which contains 'chrome' in it's name
listOfProcessIds = findProcessIdByName('chrome')
if len(listOfProcessIds) > 0:
   print('Process Exists | PID and other details are')
   for elem in listOfProcessIds:
       processID = elem['pid']
       processName = elem['name']
       processCreationTime =  time.strftime('%Y-%m-%d %H:%M:%S', time.localtime(elem['create_time']))
       print((processID ,processName,processCreationTime ))
else :
   print('No Running Process found with this text')

Now, the contents of the list

(2604, 'chrome.exe', '2018-11-10 19:12:13')
(4276, 'chrome.exe', '2018-11-10 19:12:14')
(9136, 'chrome.exe', '2018-11-10 19:12:14')
(9616, 'chrome.exe', '2018-11-10 19:43:41')
(12904, 'chrome.exe', '2018-11-10 19:12:13')
(13476, 'chrome.exe', '2018-11-10 20:03:04')
(15520, 'chrome.exe', '2018-11-10 20:02:22')

This can also be done using a single line using list comprehension  i.e

# Finding the PIDs of all the running instances of process that contains 'chrome' in it's name
procObjList = [procObj for procObj in psutil.process_iter() if 'chrome' in procObj.name().lower() ]

And the same output will come like

psutil.Process(pid=2604, name='chrome.exe', started='19:12:13')
psutil.Process(pid=4276, name='chrome.exe', started='19:12:14')
psutil.Process(pid=9136, name='chrome.exe', started='19:12:14')
psutil.Process(pid=9616, name='chrome.exe', started='19:43:41')
psutil.Process(pid=12904, name='chrome.exe', started='19:12:13')
psutil.Process(pid=13476, name='chrome.exe', started='20:03:04')
psutil.Process(pid=15520, name='chrome.exe', started='20:02:22')

Related Programs:

• Python Program to Map Two Lists into a Dictionary
• Python Program to Remove Adjacent Duplicate Characters from a String
• Python Check if Two Lists are Equal
• Python Program to Merge two Lists and Sort it
• Python Program for Multiplication of Two Matrices

Placement new operator and it’s need

This article is all about placement new operator.

• Placement new Operator
• Need of placement new operator

Placement new operator :

placement new operator we use to pass a memory address to new as a parameter. This memory is used by the placement new operator to create the object and also to call the constructor on it and then returning the same passed address.

Need of placement new operator :

As we know when we create any object using new operator, then the memory is allocated on the heap.

int * ptr = new int;

But, while working sometimes we need to create an object dynamically for which some specific memory location will be allocated.

For example, we do not want new memory to be allocated on heap rather it needs to be allocated on a given memory address. Actually this scenario comes when we work on any embedded product or with shared memory. So, for this requirement we use placement new operator.

Below is an example code to achieve this :

// Program

#include <iostream>
#include <cstdlib>
#include <new>

int main()
{
// Here memory will not be allocated on heap.
int * space = new int[1004];
// It will use passed spacer to allocate the memory
int * ptr = new(space) int;
*ptr = 7;
std::cout<<(*ptr)<<std::endl;
delete [] buffer;
return 0;
}

Create a Numpy array of evenly spaced samples over a range using numpy.linspace()

In this article we will discus about how to create a Numpy array of evenly spaced samples over a range using numpy.linspace().

Numpy.linspace()

Numpy module provides a function numpy.linspace() to create a evenly spaced samples over a specified interval i.e.

Syntax: numpy.linspace(start, stop, num=50, endpoint=True, retstep=False, dtype=None)

• start : Represents the start value of the range
• stop : Represents the end value of the range (Actually array doesn’t include this value but it acts as an end point)
• num : Represents the number of samples to be generated (If it is not provided then optional value is taken as 50)
• dtype : Represents datatype of elements.

To use Numpy we have to include the below module i.e.

import numpy as np

Let’s see below two examples to understand the concept.

• Create 5 evenly spaced samples in interval [10, 50]
• Get the Step size from numpy.linspace()

Example-1 : Create 5 evenly spaced samples in interval [10, 50]

import numpy as np
# Create 5 evenly spaced samples in interval [10, 50]
temp_arr = np.linspace(10,50,5)
print(temp_arr)

Output :
[10. 20. 30. 40. 50.]

import numpy as np
# Create 5 evenly spaced samples in interval [10, 50]
temp_arr = np.linspace(10,50,5,dtype=np.int32)
print(temp_arr)

Output :
[10 20 30 40 50]

Example-2 : Get the Step size from numpy.linspace()

When we will pass retstep=True in numpy.linspace().

import numpy as np
# Create 5 evenly spaced samples in interval [10, 50]
temp_arr, step = np.linspace(10,50,5,dtype=np.int32,retstep=True)
print('Printing the Numpy Array contents: ')
print(temp_arr)
print('The step size between two samples  : ', step)

Output :
Contents of the Numpy Array : 
[10 20 30 40 50]
Step size between two elements : 10.0