You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-user@hadoop.apache.org by 林育智 <my...@gmail.com> on 2012/06/11 09:11:37 UTC

snappy could deal with empty file in version 1.02

hi all:
    i use snappy in map output,then in hive:
    set hive.exec.compress.output=true;insert overwrite table test2 select
* from test1;

Re: snappy could deal with empty file in version 1.02

Posted by 林育智 <my...@gmail.com>.
2012/6/11  <my...@gmail.com>

> hi all:
>     i use snappy in map output,then in hive:
>     set hive.exec.compress.output=true;insert overwrite table test2 select
> * from test1;
>     the table test1 is null,then will create a empty file in test2, it's
> size is 4bytes, all zero.  when use "select * from test2" it will occur a
> eof exception. In the class BlockDecompressorStream,the decompress logical
> is :1,read the original size of chunk(4bytes);2,read size of compressed
> chunk(4bytes). but the empty file in snappy only 4 bytes.then eof exception
> occur in the step 2.  any one could help me,thx.
>

Re: snappy could deal with empty file in version 1.02

Posted by 林育智 <my...@gmail.com>.
2012/6/11  <my...@gmail.com>

> hi all:
>     i use snappy in map output,then in hive:
>     set hive.exec.compress.output=true;insert overwrite table test2 select
> * from test1;
>     the table test1 is null,then will create a empty file in test2, it's
> size is 4bytes, all zero.  when use "select * from test2" it will occur a
> eof exception. In the class BlockDecompressorStream,the decompress logical
> is :1,read the original size of chunk(4bytes);2,read size of compressed
> chunk(4bytes). but the empty file in snappy only 4 bytes.then eof exception
> occur in the step 2.  any one could help me,thx.
>