Writing a Python list of lists to a csv file

Question

I have a long list of lists of the following form ---

a = [[1.2,'abc',3],[1.2,'werew',4],........,[1.4,'qew',2]]

i.e. the values in the list are of different types -- float,int, strings.How do I write it into a csv file so that my output csv file looks like

1.2,abc,3
1.2,werew,4
.
.
.
1.4,qew,2

Solution 1

Python's built-in CSV module can handle this easily:

import csv

with open("output.csv", "wb") as f:
    writer = csv.writer(f)
    writer.writerows(a)

This assumes your list is defined as a, as it is in your question. You can tweak the exact format of the output CSV via the various optional parameters to csv.writer() as documented in the library reference page linked above.

Update for Python 3

import csv

with open("out.csv", "w", newline="") as f:
    writer = csv.writer(f)
    writer.writerows(a)

Solution 2

You could use pandas:

In [1]: import pandas as pd

In [2]: a = [[1.2,'abc',3],[1.2,'werew',4],[1.4,'qew',2]]

In [3]: my_df = pd.DataFrame(a)

In [4]: my_df.to_csv('my_csv.csv', index=False, header=False)

Solution 3

import csv
with open(file_path, 'a') as outcsv:   
    #configure writer to write standard csv file
    writer = csv.writer(outcsv, delimiter=',', quotechar='|', quoting=csv.QUOTE_MINIMAL, lineterminator='\n')
    writer.writerow(['number', 'text', 'number'])
    for item in list:
        #Write item to outcsv
        writer.writerow([item[0], item[1], item[2]])

official docs: http://docs.python.org/2/library/csv.html

Solution 4

Using csv.writer in my very large list took quite a time. I decided to use pandas, it was faster and more easy to control and understand:

 import pandas

 yourlist = [[...],...,[...]]
 pd = pandas.DataFrame(yourlist)
 pd.to_csv("mylist.csv")

The good part you can change somethings to make a better csv file:

 yourlist = [[...],...,[...]]
 columns = ["abcd","bcde","cdef"] #a csv with 3 columns
 index = [i[0] for i in yourlist] #first element of every list in yourlist
 not_index_list = [i[1:] for i in yourlist]
 pd = pandas.DataFrame(not_index_list, columns = columns, index = index)

 #Now you have a csv with columns and index:
 pd.to_csv("mylist.csv")

Solution 5

If for whatever reason you wanted to do it manually (without using a module like csv,pandas,numpy etc.):

with open('myfile.csv','w') as f:
    for sublist in mylist:
        for item in sublist:
            f.write(item + ',')
        f.write('\n')

Of course, rolling your own version can be error-prone and inefficient ... that's usually why there's a module for that. But sometimes writing your own can help you understand how they work, and sometimes it's just easier.

View more solutions

Comments

[Burhan Khalid]

This would get my +1 if you could explain your answer with some comments.

[Amber]

writerow does not take multiple arguments.

[Amber]

>>> w.writerow("a", "b", "c") Traceback (most recent call last): File "<stdin>", line 1, in <module> TypeError: writerow() takes exactly one argument (3 given)

[Dmitry Zagorulkin]

@Amber what version of python do you use?

[Dmitry Zagorulkin]

@Amber excuse me. i missed []

[Vlad V]

For Python 3 compatibility, remove the "b" from "wb".

[Spas]

With Python 3 - open('output.csv', 'w', newline=''). I get an extra line if I omit the newline parameter. docs.python.org/3/library/csv.html#csv.writer

[Tim Mottram]

In python3 I had to use open('output.csv', 'w', newline="")

[Cloud]

I don't think should use pandas if built-in library csv can do it.

[dorbodwolf]

i like pandas because its powerful

[Rambatino]

WOW that python 3 error is very unhelpful. Thanks @vladV (a bytes-like object is required, not 'str'). It kinda makes sense in hindsight, but not informative of where to look at all.

[SDG]

@VladV I think your comment should be mentioned on Amber's answer.

[Admin]

pandas is powerful, sure, but I'm not going to use a McLaren to drive to the corner store next door.

[tlalco]

If you get error "TypeError: 'newline' is an invalid keyword argument for this function" just remove the newline argument. with open(file_name, "w") as f:

[Amber]

@tlalco that probably means you're using Python 2, in which case you should use the first code block rather than the second. (It also means you should consider switching to Python 3.)

[anjandash]

Note: For cases, where the list of lists is very large, sequential csv write operations can make the code significantly slower. In such cases using pandas would be wiser.