If your files have .gz as extension, they will split.
> -----Original Message-----
> From: Rui Shi [mailto:[hidden email]]
> Sent: Thursday, December 13, 2007 2:53 PM
> To: [hidden email] > Subject: How to ask hadoop not to split the input
> My input is a bunch of gz files on local file system. I don't want
> to split them for mappers. How should I specify that?
I guess that the problem is that I wrote my own LineReader. In this case, the corresponding InputFormat has to specify that the input is not splitable by overriding the isSplitable() method. I have got that fixed.
----- Original Message ----
From: Owen O'Malley <[hidden email]>
To: [hidden email] Sent: Thursday, December 13, 2007 3:19:58 PM
Subject: Re: How to ask hadoop not to split the input
On Dec 13, 2007, at 3:03 PM, Runping Qi wrote:
> If your files have .gz as extension, they will split.