AttributeError when trying to use seek() to get last row of csv file

Question

I am trying to return the last row from a csv file. I am modifying another function that I wrote previously that returns the last line from a text file. It seemed to work as expected at first, but now when I call the function it throws an error.

reader.seek(0, os.SEEK_END)

AttributeError: '_csv.reader' object has no attribute 'seek'

import os
import csv
def getLastFile(filename):
    distance = 1024
    with open(filename,'rb') as f:
        reader = csv.reader(f)
        reader.seek(0, os.SEEK_END)
        if reader.tell() < distance:
            reader.seek(0, os.SEEK_SET)
            lines = reader.readlines()
            lastline = lines[-1]
        else:
            reader.seek(-1 * distance, os.SEEK_END)
            lines = reader.readlines()
            lastline = lines[-1]

    return lastline

Can someone please help me modify my code? I was pretty sure you could use seek in this way, maybe I'm mistaken?

You can `seek` and `tell` on `file` objects, but not on `csv.reader` objects as far as I can tell. I'm not sure there is a good way to do this in conjunction with `csv.reader`. — Tim Pierce, Nov 30 '13 at 05:46
I just made a small .csv file with 20 or so rows and 5 columns filled with random numbers. The code (I believe) should work with a generic .csv file, returning the last row, so I'm not sure it would help if I posted it, but I could if you want. Thanks! — Thetravellingfool, Nov 30 '13 at 05:47
I don't think this is a duplicate - the target refers to "How do I tell I'm on the last line" - this question is "How do I get the last line"... they're not synonymous... — Jon Clements, Dec 04 '13 at 17:22
Thetravellingfool: I too think **Jon** is correct and have voted to reopen—not that it matters much and even though it's now years later. `;-)` — martineau, May 08 '17 at 15:28

martineau · Accepted Answer · 2018-11-11T15:25:08.807

Here's a slight variation of the core concept in the accepted answer to the question Have csv.reader tell when it is on the last line applied to your variation of the problem. Since each row is potentially a different length, there's really no way around having to read the whole file.

import csv

def get_last_row(csv_filename):
    with open(csv_filename, 'r') as f:
        lastrow = None
        for lastrow in csv.reader(f): pass
        return lastrow

Update

Here's a simpler and likely faster way to do it using a collections.deque. I got the idea from one of the answers to the question How to read an output line containing a list of integers produced.

from collections import deque
import csv

def get_last_row(csv_filename):
    with open(csv_filename, 'r') as f:
        try:
            lastrow = deque(csv.reader(f), 1)[0]
        except IndexError:  # empty file
            lastrow = None
        return lastrow

score 0 · Answer 2 · answered Nov 30 '13 at 06:38

csv reader not support seek, so you can get csv file last line as text file, then process last line as csv. In your code, if last line length > distance, then will get only part of last line.

import os

def get_last_line(fin):
    line_len = 80
    fin.seek(0, os.SEEK_END)
    file_size = fin.tell()
    while True:
        line_len = min(line_len * 2, file_size)
        fin.seek(-line_len, os.SEEK_END)
        lines = f.readlines()
        if len(lines) > 1 or line_len == file_size:
            return lines[-1]

then read csv:

import csv

print ', '.join(csv.reader([last_line]).next())

AttributeError when trying to use seek() to get last row of csv file

2 Answers2

Linked