Python - CSV

CSV (stands for comma separated values) format is a commonly used data format used by spreadsheets and databases. The csv module in Python’s standard library presents classes and methods to perform read/write file operations in CSV format .

writer():

This function in csv module returns a writer object that converts data into a delimited string and stores in a file object. The function needs a file object created with open() function and with write permission as a parameter. Every row written in the file issues a newline character by default. To prevent additional line between rows, newline parameter is set to ''.

The writer() function by default uses 'excel' dialect. Alternate dialect parameter can be specified if required. The function also allows additional formatting parameters to be specified.

To start writing CSV file create the writer class using following statement:

>>> import csv
>>> csvfile=open(file.csv','w', newline='')
>>> obj=csv.writer(csvfile)

The writer class has following methods:

writerow():

This function writes items in a sequence (list, tuple or string) separating them by comma character.

writerows():

This function writes each sequence in a list as a comma separated line of items in the file.

Here is an example of writer() function. First parameter to the function is a file opened in ‘w’ mode. A list of tuples is then written to file using writerow() method.

>>> import csv
>>> marks=[('Seema',22,45),('Anil',21,56),('Mike',20,60)]
>>> csvfile=open(marks.csv','w', newline='')
>>> obj=csv.writer(csvfile)
>>> for row in marks:
obj.writerow(row)
>>> csvfile.close()

This will create ‘marks.csv’ file in current directory. Open it with any text editor. It will show following contents: 

Seema,22,45
Anil,21,56
Mike,20,60

Instead of iterating over the list we could also have used writerows() method. 

>>> csvfile=open(marks.csv','w', newline='')
>>> obj=csv.writer(csvfile)
>>> obj.writerows(marks)
>>> obj.close()

reader():

This function returns a reader object which is an iterator of lines in the csv file. We can use a for loop to display lines in the file. The file should be opened in 'r' mode.

>>> csvfile=open(marks.csv','r', newline='')
>>> obj=csv.reader(csvfile)
>>> for row in obj:
print (row)

['Seema', '22', '45']
['Anil', '21', '56']
['Mike', '20', '60']

Since reader object is an iterator stream, built-in next() function is also useful to display all lines in csv file. 

>>> csvfile=open(marks.csv','r', newline='')
>>> obj=csv.reader(csvfile)
>>> while True:
try:
row=next(obj)
print (row)
except StopIteration:
break

DictWriter():

This function creates a DictWriter object which is like a regular writer but maps dictionaries onto output rows. The function takes fieldnames parameter which is a sequence of keys. The file should be having write permission enabled. Since Python’s dict objects are not ordered, there is not enough information available to deduce the order in which the row should be written to file.

The DictWriter object has following method (in addition to writerow() and writerows() methods):

writeheader():

This method writes list of keys in dictionary as a comma separated line as first line in the file.

In following example, a list of dictionary items is defined. Each item in the list is a dictionary. Using writrows() method, they are written to file in comma separated manner.

>>> marks=[{'name':'Seema', 'age':22, 'marks':45}, {'name':'Anil', 'age':21, 'marks':56}, {'name':'Mike', 'age':20, 'marks':60}]
>>> csvfile=open(marks.csv','w', newline='')
>>> fields=list(marks[0].keys())
>>> obj=csv.DictWriter(csvfile, fieldnames=fields)      
>>> obj.writeheader()      
>>> obj.writerows(marks)      
>>> csvfile.close()  

The file shows following contents:

name,age,marks
Seema,22,45
Anil,21,56
Mike,20,60

DictReader():

This function returns a DictReader object from the underlying CSV file. Contents of the file can now be retrieved.

>>> csvfile=open(marks.csv','r', newline='')      
>>> obj=csv.DictReader(csvfile)

The DictReader class provides fieldnames attribute. It returns the dictionary keys used as header of file.

>>> obj.fieldnames      
['name', 'age', 'marks']

Use loop over the DictReader object to fetch individual dictionary objects

>>> for row in obj:
     print (row)

This results in following output:

OrderedDict([('name', 'Seema'), ('age', '22'), ('marks', '45')])
OrderedDict([('name', 'Anil'), ('age', '21'), ('marks', '56')])
OrderedDict([('name', 'Mike'), ('age', '20'), ('marks', '60')])

To convert OrderedDict object to normal dictionary, we have to first import OrderedDict from collections module.

>>> from collections import OrderedDict       
>>> r=OrderedDict([('name', 'Seema'), ('age', '22'), ('marks', '45')])       
>>> dict(r)       
{'name': 'Seema', 'age': '22', 'marks': '45'}

Dialect class

The csv module also defines a dialect class. Dialect is set of standards used to implement CSV protocol. The list of dialects available can be obtained by list_dialects() function

>>> csv.list_dialects()
['excel', 'excel-tab', 'unix']

 Dialect objects support following attributes:

To register new dialect type, use register_dialect() function as shown:

csv.register_dialect('unixpwd', delimiter=':', quoting=csv.QUOTE_NONE)

The csv module defines the following constants: