Path: csiph.com!fu-berlin.de!uni-berlin.de!not-for-mail
From: Ian Kelly <ian.g.kelly@gmail.com>
Newsgroups: comp.lang.python
Subject: Re: Drowning in a teacup?
Date: Fri, 1 Apr 2016 14:52:10 -0600
Lines: 27
Message-ID: <mailman.333.1459543979.28225.python-list@python.org>
References: <ndmlj2$pb0$1@gioia.aioe.org> <ndmmav$8n2$1@dont-email.me>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
In-Reply-To: <ndmmav$8n2$1@dont-email.me>
Precedence: list
Xref: csiph.com comp.lang.python:106263

On Fri, Apr 1, 2016 at 2:39 PM, Rob Gaddi
<rgaddi@highlandtechnology.invalid> wrote:
> Fillmore wrote:
>> Nope, wrong! contrary to what I thought I had understood about how
>> parameters are passed in Python, the function is acting on a copy(!) and
>> my original list is unchanged.
>>
>
> Nope, that's not your problem.  Your problem is the line:
>
>>              mylist = [mylist[i]] + mylist[:i] + mylist[i+1:]
>
> Which should be read as: "Take mylist[i], all of mylist before i,
> and all of mylist after i and create a new list from it. Store that
> new list in a variable called mylist, overwriting the previous value."
>
> If instead you were using the .insert and .remove methods, you'd be
> manipulating the existing list instead.  But you don't want to do that,
> because those methods are O(N) on the list length; manipulating the
> middle of lists is slow.

Or use slice assignment. This should work in place of the above:

    mylist[:] = [mylist[i]] + mylist[:i] + mylist[i+1:]

Still O(n), but so will be any implementation that removes something
from an arbitrary list position and inserts it at the front.