[ 
https://issues.apache.org/jira/browse/PHOENIX-7022?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17760048#comment-17760048
 ] 

ASF GitHub Bot commented on PHOENIX-7022:
-----------------------------------------

lokiore commented on code in PR #1660:
URL: https://github.com/apache/phoenix/pull/1660#discussion_r1309141121


##########
phoenix-core/src/main/java/org/apache/phoenix/query/ConnectionQueryServicesImpl.java:
##########
@@ -4186,19 +4185,30 @@ protected PhoenixConnection 
upgradeSystemCatalogIfRequired(PhoenixConnection met
         }
         if (currentServerSideTableTimeStamp < 
MIN_SYSTEM_TABLE_TIMESTAMP_5_2_0) {
             metaConnection = addColumnsIfNotExists(metaConnection,
-                    PhoenixDatabaseMetaData.SYSTEM_CATALOG, 
MIN_SYSTEM_TABLE_TIMESTAMP_5_2_0 -3,
+                    PhoenixDatabaseMetaData.SYSTEM_CATALOG, 
MIN_SYSTEM_TABLE_TIMESTAMP_5_2_0 - 5,
                     PhoenixDatabaseMetaData.PHYSICAL_TABLE_NAME
                             + " " + PVarchar.INSTANCE.getSqlTypeName());
 
             metaConnection = addColumnsIfNotExists(metaConnection, 
PhoenixDatabaseMetaData.SYSTEM_CATALOG,
-                MIN_SYSTEM_TABLE_TIMESTAMP_5_2_0 -2,
+                MIN_SYSTEM_TABLE_TIMESTAMP_5_2_0 - 4,
                     PhoenixDatabaseMetaData.SCHEMA_VERSION + " " + 
PVarchar.INSTANCE.getSqlTypeName());
             metaConnection = addColumnsIfNotExists(metaConnection, 
PhoenixDatabaseMetaData.SYSTEM_CATALOG,
-                MIN_SYSTEM_TABLE_TIMESTAMP_5_2_0 -1,
+                MIN_SYSTEM_TABLE_TIMESTAMP_5_2_0 - 3,
                 PhoenixDatabaseMetaData.EXTERNAL_SCHEMA_ID + " " + 
PVarchar.INSTANCE.getSqlTypeName());
             metaConnection = addColumnsIfNotExists(metaConnection, 
PhoenixDatabaseMetaData.SYSTEM_CATALOG,
-                MIN_SYSTEM_TABLE_TIMESTAMP_5_2_0,
+                MIN_SYSTEM_TABLE_TIMESTAMP_5_2_0 - 2,
                 PhoenixDatabaseMetaData.STREAMING_TOPIC_NAME + " " + 
PVarchar.INSTANCE.getSqlTypeName());
+            /**
+             * TODO: Provide a path to copy existing data from PHOENIX_TTL to 
TTL column and then
+             * to DROP PHOENIX_TTL Column. See PHOENIX-7023

Review Comment:
   This is going into a feature branch so 7022 and 7023 will go in one commit 
into master, after this feature will need a release stating that after that 
release we are going to support phoenix level ttl as TTL (DDL) only and not as 
PHOENIX_TTL (DDL).





> Add new columns TTL and ROW_KEY_PREFIX
> --------------------------------------
>
>                 Key: PHOENIX-7022
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-7022
>             Project: Phoenix
>          Issue Type: Sub-task
>            Reporter: Jacob Isaac
>            Assignee: Lokesh Khurana
>            Priority: Major
>
> When a view statement is defined by the constraints articulated in 
> PHOENIX-4555, all rows created by the view will be prefixed by a KeyRange. 
> The view thus can simply be represented by the prefixed KeyRange generated by 
> the expression representing the view statement. In other words, there exists 
> a one-to-one mapping between the view (defined by tenant, schema, tablename) 
> and PREFIXED KeyRange.
> For lookup on the PREFIXED KeyRange we will create a new column 
> ROW_KEY_PREFIX in SYSTEM.CATALOG. This new column will be populated during 
> view creation when TTL is specified.
>  
> The TTL column (INTEGER) will store the TTL when specified in line with the 
> HBase spec (which uses an int). The PHOENIX_TTL-related columns and code will 
> be deprecated in a separate jira.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to