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

Switch to Threaded View
Sqoop >> mail # user >> Mapping of relational tables and join keys to HBase

Copy link to this message
Re: Mapping of relational tables and join keys to HBase
Yes, I have checked. But I thought there was some automatic way of sqoop
doing that (without users have to explicitly perform the import for each
particular table/column).

I thought Sqoop could transform relational schemas into hbase "schemas",
i.e., to get a mapping from one structure to another. But this is not the
case right?

On Wed, Dec 18, 2013 at 9:01 PM, Abraham Elmahrek <[EMAIL PROTECTED]> wrote:

> Hey there,
> Have you checked out
> http://sqoop.apache.org/docs/1.4.4/SqoopUserGuide.html#_importing_data_into_hbase? The row key will be what ever column you are splitting by, unless
> --hbase-row-key option is specified. With a join, this should be no
> different. The resulting columns of the join will end up in HBase under the
> same column family using column names for columns in Hbase.
> -Abe
> On Wed, Dec 18, 2013 at 10:08 AM, Saiph Kappa <[EMAIL PROTECTED]>wrote:
>> Hi,
>> Can anyone tell me how sqoop maps relational keys to hbase keys?
>> For instance, in mysql, if I have table Location with columns id,
>> address, city, country, where the last two are foreign keys to tables City
>> and Country, how is this mapped to HBase? Will we still have a table City
>> and Country?
>> Or, if we have tables Consumer, Provider, and Transaction (which is just
>> to join the other two tables when a transaction happens), how is this
>> situation mapped to hbase (in terms of tables and generated keys).
>> Thanks.