Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
HBase >> mail # user >> Sequential read-update HFile

Copy link to this message
Re: Sequential read-update HFile

Once a StoreFile is written to disk it is never updated.


So the options are are either Puts (like you've been describing - which
will create new StoreFiles per MemStore flush) or bulk loading HFiles (new
Out of curiosity, are you trying to update *every* row?  Most of the rows?
On 5/8/12 12:04 PM, "Nikolaos Papailiou" <[EMAIL PROTECTED]> wrote:

>Hi all,
>I would like to sequentially read and update some HFiles. That means
>that I want something like a HFileScanner but with the capability of
>directly updating the current key/value pair. I have been trying to do
>that with sequential puts and major compactions of the region but the
>performance is very poor. Is there another way to do this efficiently?