Hi tariq ,

   Have a look on this link which can guide you ..
There was discussion happen previously for the same type of issue

search-hadoop.com/m/ydCoSysmTd1

Syed Abdul kather
send from Samsung S3
On Aug 6, 2012 11:48 PM, "Manoj Khangaonkar" <[email protected]> wrote:

> Hi,
>
> I think you might need to extend FileInputFormat ( or one of its
> derived classes)  as well as
> implement a RecordReader.
>
> regards
>
> On Mon, Aug 6, 2012 at 8:30 AM, Mohammad Tariq <[email protected]> wrote:
> > Hello list,
> >
> >      I need some guidance on how to handle files where we don't have
> > any proper delimiters or record boundaries. Actually I am trying to
> > process a set of file that are totally alien to me (SAS XPT files)
> > through MR. But one thing that is always fixed is that each time I
> > have to read 107 bytes from the line. Is it possible to use this
> > length as a delimiter for creating splits some how??And if so which
> > InputFormat would be appropriate??Many thanks.
> >
> > Regards,
> >     Mohammad Tariq
>
>
>
> --
> http://khangaonkar.blogspot.com/
>

Reply via email to