Python: No csv.close()?

Question

The reader is really just a parser. When you ask it for a line of data, it delegates the reading action to the underlying file object and just converts the result into a set of fields. The reader itself doesn't manage any resources that would need to be cleaned up when you're done using it, so there's no need to close it; it'd be a meaningless operation.

You should make sure to close the underlying file object, though, because that does manage a resource (an open file descriptor) that needs to be cleaned up. Here's the way to do that:

with open('/home/rv/ncbi-blast-2.2.23+/db/output.blast') as f:
    z = csv.reader(f, delimiter='\t')
    # do whatever you need to with z

If you're not familiar with the with statement, it's roughly equivalent to enclosing its contents in a try...finally block that closes the file in the finally part.

Hopefully this doesn't matter (and if it does, you really need to update to a newer version of Python), but the with statement was introduced in Python 2.5, and in that version you would have needed a __future__ import to enable it. If you were working with an even older version of Python, you would have had to do the closing yourself using try...finally.

---

^{Thanks to Jared for pointing out compatibility issues with the with statement.}

You do not close CSV readers directly; instead you should close whatever file-like object is being used. For example, in your case, you'd say:

f = open('/home/rv/ncbi-blast-2.2.23+/db/output.blast')
z = csv.reader(f, delimiter='\t')
...
f.close()

If you are using a recent version of Python, you can use the with statement, e.g.

with open('/home/rv/ncbi-blast-2.2.23+/db/output.blast') as f:
    z = csv.reader(f, delimiter='\t')
    ...

This has the advantage that f will be closed even if you throw an exception or otherwise return inside the with-block, whereas such a case would lead to the file remaining open in the previous example. In other words, it's basically equivalent to a try/finally block, e.g.

f = open('/home/rv/ncbi-blast-2.2.23+/db/output.blast')
try:
    z = csv.reader(f, delimiter='\t')
    ...
finally:
    f.close()

You don't close the result of the reader() method, you close the result of the open() method. So, use two statements: foo=open(...); bar=csv.reader(foo). Then you can call foo.close().

There are no bonus points awarded for doing in one line that which can be more readable and functional in two.