Return-Path: X-Original-To: apmail-hive-dev-archive@www.apache.org Delivered-To: apmail-hive-dev-archive@www.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id 8F9E31124C for ; Mon, 23 Jun 2014 18:07:27 +0000 (UTC) Received: (qmail 98141 invoked by uid 500); 23 Jun 2014 18:07:26 -0000 Delivered-To: apmail-hive-dev-archive@hive.apache.org Received: (qmail 98072 invoked by uid 500); 23 Jun 2014 18:07:26 -0000 Mailing-List: contact dev-help@hive.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: dev@hive.apache.org Delivered-To: mailing list dev@hive.apache.org Received: (qmail 98054 invoked by uid 500); 23 Jun 2014 18:07:26 -0000 Delivered-To: apmail-hadoop-hive-dev@hadoop.apache.org Received: (qmail 98051 invoked by uid 99); 23 Jun 2014 18:07:26 -0000 Received: from arcas.apache.org (HELO arcas.apache.org) (140.211.11.28) by apache.org (qpsmtpd/0.29) with ESMTP; Mon, 23 Jun 2014 18:07:26 +0000 Date: Mon, 23 Jun 2014 18:07:26 +0000 (UTC) From: "Laljo John Pullokkaran (JIRA)" To: hive-dev@hadoop.apache.org Message-ID: In-Reply-To: References: Subject: [jira] [Commented] (HIVE-5775) Introduce Cost Based Optimizer to Hive MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable X-JIRA-FingerPrint: 30527f35849b9dde25b450d4833f0394 [ https://issues.apache.org/jira/browse/HIVE-5775?page=3Dcom.atlassian.= jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D14041= 053#comment-14041053 ]=20 Laljo John Pullokkaran commented on HIVE-5775: ---------------------------------------------- Following may help in reducing the confusion: 1. In design doc the cost formula is for choosing Join Algorithm. The cost = formula as described in the doc assumes Tez execution. 2. However current work on CBO doesn=E2=80=99t include Join algorithm selec= tion. Instead it rearranges Join based on Join cardinality & NDV. In other = words Join reordering is not depended on Physical Execution Layer (Tez or M= R). 3. When we decide to do Join Algorithm Selection we can fit in cost formula= for both a) MR b) Tez. This way, based on the physical execution layer we = can select best Join Algorithm/Order.=20 4. The cost formula for Join Algorithm selection is not that different betw= een MR & Tez (except for intermediate HDFS writes). So assume that CBO can = support both execution layers rather easily. 5. CBO framework allows you to plug and play any cost model. There is no ha= rd coupling. > Introduce Cost Based Optimizer to Hive > -------------------------------------- > > Key: HIVE-5775 > URL: https://issues.apache.org/jira/browse/HIVE-5775 > Project: Hive > Issue Type: New Feature > Reporter: Laljo John Pullokkaran > Assignee: Laljo John Pullokkaran > Attachments: CBO-2.pdf, HIVE-5775.1.patch > > -- This message was sent by Atlassian JIRA (v6.2#6252)