1 hive instance can span multiple dfs. So, it makes sense to have fully qualified name
On 2/4/11 6:55 AM, "Chinna" <chinna...@huawei.com> wrote: >Thanks Namit, > >Here I'm interested to know like can we use full path excluding the Name >Node IP and Port. If we do like this we can have a portability like later >NameNode IP or Port got changed also we can retrieve the data. > >For Example: >Using hive with the default configurations of hadoop created a sample >table. >Later changed the NameNode port to some other port and restarted >hadoop(9000 >to 9090). In this case we are not able to access the tables created before >changing the NameNode Port because for that table path contains old port >(before changing the port). > >If we store the location of the table by excluding NameNode IP & Port and >in >hive construct full HDFS URL Path. > >Pls give u r inputs > > > > >************************************************************************** >** >*********** > > This e-mail and attachments contain confidential information >from HUAWEI, which is intended only for the person or entity whose address >is listed above. Any use of the information contained herein in any way >(including, but not limited to, total or partial disclosure, reproduction, >or dissemination) by persons other than the intended recipient's) is >prohibited. If you receive this e-mail in error, please notify the sender >by >phone or email immediately and delete it! > > >-----Original Message----- >From: Namit Jain [mailto:nj...@fb.com] >Sent: Friday, February 04, 2011 12:58 AM >To: dev@hive.apache.org; chinna...@huawei.com; hive-...@hadoop.apache.org >Subject: Re: Query regarding Metastore(Derby) SDS table data > >Different partitions can have different paths - >A partition's path need not be a sub-direcotry of the table's path. >In facebook, we use this regularly, specially for external tables. > >So, it simplifies things if the full path is stored for the partition in >the metastore. > > >Thanks, >-namit > > >On 2/3/11 2:22 AM, "Chinna" <chinna...@huawei.com> wrote: > >>Hi all, >> >> While creating hive table using create table command, the code flow will >>also insert the HDFS location in SDS table present in Derby >> >>For Example on executing:- >>create table sample(rate int) stored as textfile; >> >> >>The SDS table (meta table present in Derby) contains following entry >>corresponding to Hive table sample like >>http://{IP}:9001/user/hive/warehouse/sample >> >>say 9001 is the port configured, >>and {HDFS-URL}/user/hive/warehouse is the warehouse configured. >> >>Now here I'm interested to know the intention of maintaining full path of >>the HDFS location in SDS Table, say why not a relative path, anyhow Hive >>is >>capable of constructing full HDFS URL Path. >> >>Any design inputs on it, will surely help. >> >>Thanks, >>Chinna Rao Lalam >> >> >> >> >> >> >>************************************************************************* >>* >>** >>*********** >> >> This e-mail and attachments contain confidential information >>from HUAWEI, which is intended only for the person or entity whose >>address >>is listed above. Any use of the information contained herein in any way >>(including, but not limited to, total or partial disclosure, >>reproduction, >>or dissemination) by persons other than the intended recipient's) is >>prohibited. If you receive this e-mail in error, please notify the sender >>by >>phone or email immediately and delete it! >> >> >
