[
https://issues.apache.org/jira/browse/CARBONDATA-1518?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ravindra Pesala updated CARBONDATA-1518:
----------------------------------------
Description:
User can give timeseries option while creating the main table itself and carbon
will create aggregate tables automatically.
{code}
CREATE TABLE agg_sales
STORED BY 'carbondata'
TBLPROPERTIES ('parent_table'='sales', ‘timeseries_column’=’order_time’,
‘granualarity’=’hour’, ‘rollup’ =’quantity:sum, max # user_id: count # price:
sum, max, min, avg’)
{code}
In the above case, user choose timeseries_column, granularity and aggregation
types for measures, so carbon generates the aggregation tables automatically
for year, month, day and hour level aggregation tables (totally 4 tables, their
table name will be prefixed with agg_sales).
was:
User can give timeseries option while creating the main table itself and carbon
will create aggregate tables automatically.
{code}
create table sales (order_time timestamp, user_id string, order_id string,
status string, quantity int, price bigint) stored by 'carbondata'
tblproperties('timeseries_column'='order_time', granularity='hour'
'aggtype'='quanity:sum,max#user_id:count#price:sum,max,min,avg')
{code}
In the above case user created main table and chosen timeseries column,
granularity and aggregation types for measures, so carbon generates the
aggregation tables automatically for year, month, day and hour level
aggregation tables.
> 2. Support creating timeseries while creating main table.
> ---------------------------------------------------------
>
> Key: CARBONDATA-1518
> URL: https://issues.apache.org/jira/browse/CARBONDATA-1518
> Project: CarbonData
> Issue Type: Sub-task
> Reporter: Ravindra Pesala
>
> User can give timeseries option while creating the main table itself and
> carbon will create aggregate tables automatically.
> {code}
> CREATE TABLE agg_sales
> STORED BY 'carbondata'
> TBLPROPERTIES ('parent_table'='sales', ‘timeseries_column’=’order_time’,
> ‘granualarity’=’hour’, ‘rollup’ =’quantity:sum, max # user_id: count # price:
> sum, max, min, avg’)
> {code}
> In the above case, user choose timeseries_column, granularity and aggregation
> types for measures, so carbon generates the aggregation tables automatically
> for year, month, day and hour level aggregation tables (totally 4 tables,
> their table name will be prefixed with agg_sales).
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
