You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@phoenix.apache.org by "James Taylor (JIRA)" <ji...@apache.org> on 2015/07/31 01:43:04 UTC
[jira] [Commented] (PHOENIX-2159) Grammar changes and DDL support
for surfacing native HBase timestamp
[ https://issues.apache.org/jira/browse/PHOENIX-2159?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14648505#comment-14648505 ]
James Taylor commented on PHOENIX-2159:
---------------------------------------
Can we have ROW_TIMESTAMP a kind of optional sibling with the ASC/DESC modifier so that it's part of the primary key constraint definition? For example, the following would be allowed:
{code}
CREATE TABLE T (
K1 DATE PRIMARY KEY ROW_TIMESTAMP);
CREATE TABLE T (
K1 TIME PRIMARY KEY ROW_TIMESTAMP,
V1 VARCHAR);
CREATE TABLE T (
K1 VARCHAR,
K2 BIGINT,
CONSTRAINT PK PRIMARY KEY(K1, K2 ROW_TIMESTAMP));
{code}
We could allow DATE, and BIGIINT as the type for a column marked as the ROW_TIMESTAMP.
The following wouldn't be valid:
{code}
CREATE TABLE T (
K1 DATE ROW_TIMESTAMP, -- may only be defined as part of PK definition
K2 VARCHAR);
CREATE TABLE T (
K1 DATE PRIMARY KEY ROW_TIMESTAMP, -- invalid as essentially defining PK multiple times
K2 VARCHAR,
CONSTRAINT pk PRIMARY KEY (K1, K2));
CREATE TABLE T (
K1 DATE ROW_TIMESTAMP, -- must appear in CONSTRAINT definition below just like DESC declaration
K2 VARCHAR,
CONSTRAINT pk PRIMARY KEY (K1, K2));
CREATE TABLE T (
K1 VARCHAR,
K2 VARCHAR,
CONSTRAINT PK PRIMARY KEY(K1, K2 ROW_TIMESTAMP)); -- invalid type for ROW_TIMESTAMP
{code}
> Grammar changes and DDL support for surfacing native HBase timestamp
> --------------------------------------------------------------------
>
> Key: PHOENIX-2159
> URL: https://issues.apache.org/jira/browse/PHOENIX-2159
> Project: Phoenix
> Issue Type: Sub-task
> Reporter: Samarth Jain
> Assignee: Samarth Jain
> Attachments: PHOENIX-2159.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)