Versions Compared

Old Version 15

changes.mady.by.user Josh Elser

Saved on

compared with

New Version Current

changes.mady.by.user Josh Elser

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: Added indexed table example

...

No Format
CREATE EXTERNAL TABLE countries(key string, name string, country string, country_id int)
STORED BY 'org.apache.hadoop.hive.accumulo.AccumuloStorageHandler'
WITH SERDEPROPERTIES ("accumulo.columns.mapping" = ":rowID,info:name,info:country,info:country_id");

Create an indexed table

To take advantage of indexing, Hive uses another Accumulo table is used to create a lexicographically-sorted search term index for each field allowing for very efficient exact match and bounded range searches.

No Format
CREATE TABLE company_stats (
   rowid string,
   active_entry boolean,
   num_offices tinyint,
   num_personel smallint,
   total_manhours int,
   num_shareholders bigint,
   eff_rating float,
   err_rating double,
   yearly_production decimal,
   start_date date,
   address varchar(100),
   phone char(13),
   last_update timestamp )
ROW FORMAT SERDE 'org.apache.hadoop.hive.accumulo.serde.AccumuloSerDe'
STORED BY 'org.apache.hadoop.hive.accumulo.AccumuloStorageHandler'
WITH SERDEPROPERTIES (
   "accumulo.columns.mapping" = ":rowID,a:act,a:off,a:per,a:mhs,a:shs,a:eff,a:err,a:yp,a:sd,a:addr,a:ph,a:lu”,
   "accumulo.table.name"="company_stats",
   "accumulo.indextable.name"="company_stats_idx"
 );

Acknowledgements

I would be remiss to not mention the efforts made by Brian Femiano that were the basis for this storage handler. His initial prototype for Accumulo-Hive integration was the base for this work.