Extracting specific rows and columns from a CSV file

Question

I have written a function to selectively extract data from a file. I want to be able to extract only from a certain line and only given rows.

Would convert this function into a generator improve the overhead when I need to process large files?

import itertools
import csv
def data_extraction(filename,start_line,lenght,span_start,span_end):
 with open(filename, "r") as myfile:
 file_= csv.reader(myfile, delimiter=' ') #extracts data from .txt as lines
 return (x for x in [filter(lambda a: a != '', row[span_start:span_end]) \
 for row in itertools.islice(file_, start_line, lenght)])

• Mast: Are you looking for general advice or are you just interested in making this a generator? The former is on-topic here, the latter is not. Too specific. Please take a look at the help center.
• Sorade: Well I am asking about increasing the performance of this working function. And according to the help center, generally applicable questions on code should be going to SO, and specific question on how to improve a piece of code should go here.
• Mast: "Do I want feedback about any or all facets of the code?" This means we can complain about any and all facets of your code, even if it doesn't address your specific concern for generators. If you have a problem with that, CR is not the place to be. If you're fine with that, Welcome!
• Sorade: I'm very happy with any feedback I can get . Being self-taught I have a lot to learn so I'll take any constructive criticism.

Answer 1

Use round parenthesis for generators

Also x for x in was unnecessary:

return (filter(lambda a: a != '', row[span_start:span_end]) \
 for row in itertools.islice(file_, start_line, lenght))

If you use Python 2 you should use itertools.ifilter because it returns a generator while filter returns a list.

The functions is pretty clear overall, I suggest you space your argument list as according to PEP8 conventions. Also investigate in easier to remember argument formats like f(file, line_range, inline_range) where two tuples replace 4 arguments.