[
https://issues.apache.org/jira/browse/IGNITE-4651?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vladimir Ozerov updated IGNITE-4651:
------------------------------------
Description:
This is an umbrella ticket for all tasks related to CREATE/DROP TABLE feature.
Raw design:
1) {{CREATE TABLE}} is performed using {{createCache}}, {{DROP TABLE}} is
performed using {{destroyCache}};
2) We need to extend H2 parser to support "AFFINITY KEY" attribute;
3) Caches will be created based on cache templates defined in
{{IgniteConfiguration}}. There should be several predefined templates for the
most common cases (e.g. {{PARTITIONED}}/{{REPLICATED}}). Template will be
specified using {{WITH}} feature of H2 engine.
4) Currently every cache lives inside it's own H2 schema. For user convenience
we will have to allow different caches to share the same schema.
5) Current SQL API are tightly coupled to particular cache. This leads to
anecdotal situation when you cannot call {{CREATE TABLE}} on a cluster without
caches. For this reason we will need better SQL API which is not bound to any
cache.
> Support CREATE TABLE and DROP TABLE commands
> --------------------------------------------
>
> Key: IGNITE-4651
> URL: https://issues.apache.org/jira/browse/IGNITE-4651
> Project: Ignite
> Issue Type: New Feature
> Components: SQL
> Reporter: Vladimir Ozerov
>
> This is an umbrella ticket for all tasks related to CREATE/DROP TABLE feature.
> Raw design:
> 1) {{CREATE TABLE}} is performed using {{createCache}}, {{DROP TABLE}} is
> performed using {{destroyCache}};
> 2) We need to extend H2 parser to support "AFFINITY KEY" attribute;
> 3) Caches will be created based on cache templates defined in
> {{IgniteConfiguration}}. There should be several predefined templates for the
> most common cases (e.g. {{PARTITIONED}}/{{REPLICATED}}). Template will be
> specified using {{WITH}} feature of H2 engine.
> 4) Currently every cache lives inside it's own H2 schema. For user
> convenience we will have to allow different caches to share the same schema.
> 5) Current SQL API are tightly coupled to particular cache. This leads to
> anecdotal situation when you cannot call {{CREATE TABLE}} on a cluster
> without caches. For this reason we will need better SQL API which is not
> bound to any cache.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
