Writing to a CSV file in Python

Question

I'm totally new to Python, but am pretty well-versed in other languages so I have at least some idea of what's going on. I did not write this code, but I am trying to make it work for my purposes.

I'm using an API to retrieve statistics about every play in the NFL from last season. I am trying to write this information to a CSV file, and so far, it's writing the headers to the file and not the actual data.

Can someone tell me what I'm doing wrong?

import csv 
import nflgame 

games = nflgame.games_gen(2013, kind='REG') 
plays = nflgame.combine_plays(games) 
headers = [] 
for statId, info in nflgame.statmap.idmap.iteritems(): 
    headers += info['fields'] 
    headers.sort() 
    headers = ['gsis_id', 'drive_id', 'play_id', 'desc'] + headers 
    writer = csv.DictWriter(open('2013_regular_plays.csv', 'w+'), fieldnames=headers) 
    writer.writerow({k: k for k in headers}) 
    for p in plays: 
        row = {k: getattr(p, k, 0) for k in headers} 
        row['gsis_id'] = p.drive.game.eid 
        row['drive_id'] = p.drive_num 
        row['play_id'] = p.playid 
        row['desc'] = p.desc 
        writer.writerow(row)

Answer 1

This looks like it should mostly work.

The only detail which is wrong compared to the documentation is that the file should be open in binary mode (w+b).

Also, it is important that the file be closed before you can have a look at it:

with open('2013_regular_plays.csv', 'w+b') as csv_file:
    writer = csv.DictWriter(csv_file, fieldnames=headers)
    …

will automatically close the file after the with block (and the file mode could be more simply wb, if the file is not read inside this block). If you look at the file before it is closed, its contents might still reside in RAM instead of on the disk.

PS: As DSM pointed out, each iteration for statId, info in… empties the CSV file when you open it with the w+ (or w+b) mode. If the last iteration has no plays, then the file end up being empty (with headers only). You typically want to open the CSV file before the loop (maybe even simply with the wb mode).