Return-Path: X-Original-To: apmail-avro-dev-archive@www.apache.org Delivered-To: apmail-avro-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 557D518A7D for ; Thu, 21 Jan 2016 20:44:40 +0000 (UTC) Received: (qmail 15998 invoked by uid 500); 21 Jan 2016 20:44:40 -0000 Delivered-To: apmail-avro-dev-archive@avro.apache.org Received: (qmail 15884 invoked by uid 500); 21 Jan 2016 20:44:40 -0000 Mailing-List: contact dev-help@avro.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: dev@avro.apache.org Delivered-To: mailing list dev@avro.apache.org Received: (qmail 15803 invoked by uid 99); 21 Jan 2016 20:44:40 -0000 Received: from arcas.apache.org (HELO arcas) (140.211.11.28) by apache.org (qpsmtpd/0.29) with ESMTP; Thu, 21 Jan 2016 20:44:40 +0000 Received: from arcas.apache.org (localhost [127.0.0.1]) by arcas (Postfix) with ESMTP id DED482C14F2 for ; Thu, 21 Jan 2016 20:44:39 +0000 (UTC) Date: Thu, 21 Jan 2016 20:44:39 +0000 (UTC) From: =?utf-8?Q?Cl=C3=A9ment_MATHIEU_=28JIRA=29?= To: dev@avro.apache.org Message-ID: In-Reply-To: References: Subject: [jira] [Created] (AVRO-1787) Add support of directories & globs to concat MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable X-JIRA-FingerPrint: 30527f35849b9dde25b450d4833f0394 Cl=C3=A9ment MATHIEU created AVRO-1787: ------------------------------------- Summary: Add support of directories & globs to concat Key: AVRO-1787 URL: https://issues.apache.org/jira/browse/AVRO-1787 Project: Avro Issue Type: Bug Components: java Affects Versions: 1.7.7 Reporter: Cl=C3=A9ment MATHIEU While answering [this question|http://stackoverflow.com/questions/34856838/= concat-avro-files-using-avro-tools/34899425#34899425] on stack overflow, I = noticed that the concat tool does not support directories or glob patterns = as "input-file". OP's use case seems very valid but accepting only files as input push the = burden on users. I think that it makes sense to also accept directories as input like cat do= es. Supporting glob patterns seems also useful if one want to concatenate t= he output of a MR job (in this case passing the directory won't work becaus= e of the {{_SUCESS}} file). I have quickly patched {{concat}} to support files, directories and glob pa= tterns. Support of glob patterns has been added to {{cat}} since it seems u= seful too. This change should not introduce regression for users. But any feedback abo= ut glob patterns is welcome. -- This message was sent by Atlassian JIRA (v6.3.4#6332)