[
https://issues.apache.org/jira/browse/PHOENIX-7514?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Viraj Jasani updated PHOENIX-7514:
----------------------------------
Description:
So far, we use "CREATE CDC" and "DROP CDC" grammar to create and drop multiple
CDC objects for the given table. However, we should allow creating only single
CDC object for the given table because the CDC object internally creates CDC
Index for the given table. As this index is built on PARTITION_ID() and
PHOENIX_ROW_TIMESTAMP(), having multiple similar index can be overkill for the
CDC purpose. Once the CDC Stream is created, CDC consumer can define start and
end timestamp that they are interested in.
In order to hide the details of CDC object from the user, we should define
grammar such that it allows building CDC Index only once per table. Once the
CDC object and index is created, we should not allow creating more CDC objects
for the given table.
Proposed new grammar to enable the CDC stream on the given table:
{code:java}
ENABLE CDC ON <table-name>{code}
Proposed new grammar to disable the CDC stream on the given table:
{code:java}
DISABLE CDC ON <table-name>{code}
As we no longer need to provide CDC object name with the grammar, client
inherently creates CDC object name as
_phoenix-cdc-stream-<tablename>-<timestamp>_
was:
So far, we use "CREATE CDC" and "DROP CDC" grammar to create and drop multiple
CDC objects for the given table. However, we should allow creating only single
CDC object for the given table because the CDC object internally creates CDC
Index for the given table. As this index is built on PARTITION_ID() and
PHOENIX_ROW_TIMESTAMP(), having multiple similar index can be overkill for the
CDC purpose. Once the CDC Stream is created, CDC consumer can define start and
end timestamp that they are interested in.
In order to hide the details of CDC object from the user, we should define
grammar such that it allows building CDC Index only once per table. Once the
CDC object and index is created, we should not allow creating more CDC objects
for the given table.
Proposed new grammar to enable the CDC stream on the given table:
{code:java}
ENABLE CDC ON <table-name>{code}
Proposed new grammar to disable the CDC stream on the given table:
{code:java}
DISABLE CDC ON <table-name> {code}
As we no longer need to provide CDC object name with the grammar, client
inherently creates CDC object name as
_phoenix-cdc-stream-<tablename>-<timestamp>_
> CDC grammar for enabling or disabling Stream
> --------------------------------------------
>
> Key: PHOENIX-7514
> URL: https://issues.apache.org/jira/browse/PHOENIX-7514
> Project: Phoenix
> Issue Type: Sub-task
> Reporter: Viraj Jasani
> Priority: Major
>
> So far, we use "CREATE CDC" and "DROP CDC" grammar to create and drop
> multiple CDC objects for the given table. However, we should allow creating
> only single CDC object for the given table because the CDC object internally
> creates CDC Index for the given table. As this index is built on
> PARTITION_ID() and PHOENIX_ROW_TIMESTAMP(), having multiple similar index can
> be overkill for the CDC purpose. Once the CDC Stream is created, CDC consumer
> can define start and end timestamp that they are interested in.
> In order to hide the details of CDC object from the user, we should define
> grammar such that it allows building CDC Index only once per table. Once the
> CDC object and index is created, we should not allow creating more CDC
> objects for the given table.
> Proposed new grammar to enable the CDC stream on the given table:
> {code:java}
> ENABLE CDC ON <table-name>{code}
> Proposed new grammar to disable the CDC stream on the given table:
> {code:java}
> DISABLE CDC ON <table-name>{code}
> As we no longer need to provide CDC object name with the grammar, client
> inherently creates CDC object name as
> _phoenix-cdc-stream-<tablename>-<timestamp>_
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
