Path: csiph.com!usenet.pasdenom.info!news.albasani.net!newsfeed.freenet.ag!news2.euro.net!newsgate.cistron.nl!newsgate.news.xs4all.nl!post.news.xs4all.nl!not-for-mail Return-Path: X-Original-To: python-list@python.org Delivered-To: python-list@mail.python.org X-Spam-Status: OK 0.005 X-Spam-Evidence: '*H*': 0.99; '*S*': 0.00; 'pop': 0.05; 'element': 0.07; 'indexing': 0.07; 'smallest': 0.07; 'replied': 0.09; 'subject:number': 0.09; 'cc:addr:python-list': 0.11; 'anyway': 0.14; 'dictionary,': 0.16; 'element.': 0.16; 'heap': 0.16; 'heapq': 0.16; 'inserting': 0.16; 'iterating': 0.16; 'likewise': 0.16; 'subject:Prime': 0.16; 'subject:generator': 0.16; 'swaps': 0.16; 'pushed': 0.16; 'sender:addr:gmail.com': 0.17; 'wrote:': 0.18; 'code.': 0.18; 'wed,': 0.18; 'dependent': 0.19; 'cc:addr:python.org': 0.22; 'fine': 0.24; 'cc:2**0': 0.24; 'performing': 0.26; 'push': 0.26; 'header:In-Reply-To:1': 0.27; 'am,': 0.29; 'returned': 0.30; 'message-id:@mail.gmail.com': 0.30; 'getting': 0.31; '>>>>': 0.31; "i'd": 0.34; 'problem': 0.35; 'but': 0.35; 'received:google.com': 0.35; 'module.': 0.36; 'too': 0.37; 'list': 0.37; 'implement': 0.38; 'minimum': 0.38; 'saves': 0.38; 'ian': 0.60; 'removing': 0.60; 'new': 0.61; 'matter': 0.61; 'real': 0.63; 'july': 0.63; 'to:addr:gmail.com': 0.65; 'here': 0.66; 'jul': 0.74; 'actually,': 0.84; 'hood': 0.84; 'working,': 0.84; 'imagine': 0.93; '2013': 0.98 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:sender:in-reply-to:references:from:date :x-google-sender-auth:message-id:subject:to:cc:content-type :content-transfer-encoding; bh=EtSXHW27bosKECKVAIpNRo15QtTwUpBJtQMWNiV8OqA=; b=TzonEqwd1mDDMm+k0SI4nyYdYYVGsoiNyPdo9C0dDEBENsfiK9OW6Qsbw+6kozc2X0 9uICkV03cIaJWxrRGcy4AbIC+zkcLEVFRN1Hy4UuPv2iGtEYmtlOVP2cqt9lcZ1R3q+B 1PBj5H4NBBwCcrKDytbviZ1OC+Pc4/icjtXDr7Zdls1RoCN/EYUpdfAKE/fv9OuasCzS ZKLNAIu9vj+Fcqppa9uTtsL/9IJUaKo86/j+QxTqRjiwy7VqTBLp/Dx3pW33PjetiYyQ fDp/JHQ93u5Lq78q4Y8DEBzXkU2vIpmPoQ/+ZWviZYy4Q0CzK1iC5PP5o5czBKLLKOjS 38kg== X-Received: by 10.112.42.44 with SMTP id k12mr15643911lbl.63.1373483245129; Wed, 10 Jul 2013 12:07:25 -0700 (PDT) MIME-Version: 1.0 Sender: joshua.landau.ws@gmail.com In-Reply-To: References: <15167633-b6e7-46cc-a043-8dfe8aaad11e@googlegroups.com> From: Joshua Landau Date: Wed, 10 Jul 2013 20:06:44 +0100 X-Google-Sender-Auth: i5yRtz2bNO9Gq2fr-MmOLZhI_Z0 Subject: Re: Prime number generator To: Ian Kelly Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Cc: Python X-BeenThere: python-list@python.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: General discussion list for the Python programming language List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Newsgroups: comp.lang.python Message-ID: Lines: 27 NNTP-Posting-Host: 2001:888:2000:d::a6 X-Trace: 1373483247 news.xs4all.nl 15886 [2001:888:2000:d::a6]:37871 X-Complaints-To: abuse@xs4all.nl Xref: csiph.com comp.lang.python:50389 On 10 July 2013 19:56, Ian Kelly wrote: > On Wed, Jul 10, 2013 at 11:47 AM, Joshua Landau wrote: >>>> If you care about speed, you might want to check the heapq module. Rem= oving the smallest item and inserting a new item in a heap both cost O(log(= N)) time, while finding the minimum in a dictionary requires iterating over= the whole dictionary, which cost O(N) time. >> >> Actually, because it's a list under the hood I'd imagine push and pop >> still take O(n) time :/. > > It shouldn't. You can implement push by appending the new item and > then getting it into the right place by performing O(log n) swaps. > Likewise for pop, you can update the heap with O(log n) swaps and then > removing the tail element. Appending or removing the tail element of > a list is amortized O(1). Genius. Bas replied off-list that it won't matter too much anyway as they're over-allocated, but this makes tons of sense. >> PS: It's faster to use heapreplace(...) than >> heappop(...);heappush(...) but it only saves a few %. > > The problem with heapreplace here is that the value to be pushed > is dependent on the value returned by pop. That's fine because indexing is much faster than pop. The "few %" was a real statistic from working, timed code.