[
https://issues.apache.org/jira/browse/HIVE-25813?focusedWorklogId=812759&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-812759
]
ASF GitHub Bot logged work on HIVE-25813:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 28/Sep/22 06:43
Start Date: 28/Sep/22 06:43
Worklog Time Spent: 10m
Work Description: dengzhhu653 commented on code in PR #3627:
URL: https://github.com/apache/hive/pull/3627#discussion_r982004642
##########
ql/src/java/org/apache/hadoop/hive/ql/ddl/table/create/like/CreateTableLikeOperation.java:
##########
@@ -157,27 +158,12 @@ private void setTableParameters(Table tbl) throws
HiveException {
}
// We should copy only those table parameters that are specified in the
config.
SerDeSpec spec = AnnotationUtils.getAnnotation(serdeClass,
SerDeSpec.class);
-
- // for non-native table, property storage_handler should be retained
- retainer.add(META_TABLE_STORAGE);
- if (spec != null && spec.schemaProps() != null) {
- retainer.addAll(Arrays.asList(spec.schemaProps()));
- }
-
- String paramsStr = HiveConf.getVar(context.getConf(),
HiveConf.ConfVars.DDL_CTL_PARAMETERS_WHITELIST);
- if (paramsStr != null) {
- retainer.addAll(Arrays.asList(paramsStr.split(",")));
- }
-
- Map<String, String> params = tbl.getParameters();
- if (!retainer.isEmpty()) {
- params.keySet().retainAll(retainer);
- } else {
- params.clear();
- }
+ //With Hive-25813, we'll not copy over table properties from the source.
+ //CTLT should should copy column schema but not table properties. It is
also consistent with other query engines like mysql, redshift.
+ tbl.getParameters().clear();
Review Comment:
I'm wondering this would introduce incompatibility, such as discovery
partitions, compression etc.
There is also a property `hive.ddl.createtablelike.properties.whitelist`,
which select the table Properties to copy over in this case.
Issue Time Tracking
-------------------
Worklog Id: (was: 812759)
Time Spent: 1h 40m (was: 1.5h)
> CREATE TABLE x LIKE storagehandler-based-source fails
> ------------------------------------------------------
>
> Key: HIVE-25813
> URL: https://issues.apache.org/jira/browse/HIVE-25813
> Project: Hive
> Issue Type: Bug
> Reporter: Sai Hemanth Gantasala
> Assignee: Sai Hemanth Gantasala
> Priority: Major
> Labels: pull-request-available
> Time Spent: 1h 40m
> Remaining Estimate: 0h
>
> {code:java}
> CREATE EXTERNAL TABLE default.dbs (
> DB_ID bigint,
> DB_LOCATION_URI string,
> NAME string,
> OWNER_NAME string,
> OWNER_TYPE string )
> STORED BY 'org.apache.hive.storage.jdbc.JdbcStorageHandler'
> TBLPROPERTIES (
> 'hive.sql.database.type' = 'MYSQL',
> 'hive.sql.jdbc.driver' = 'com.mysql.jdbc.Driver',
> 'hive.sql.jdbc.url' = 'jdbc:mysql://localhost:3306/hive1',
> 'hive.sql.dbcp.username' = 'hive1',
> 'hive.sql.dbcp.password' = 'cloudera',
> 'hive.sql.query' = 'SELECT DB_ID, DB_LOCATION_URI, NAME, OWNER_NAME,
> OWNER_TYPE FROM DBS'
> );
> CREATE TABLE default.dbscopy LIKE default.dbs;
> Caused by: java.lang.NullPointerException
> at
> org.apache.hadoop.hive.metastore.HiveMetaStoreUtils.getFieldsFromDeserializer(HiveMetaStoreUtils.java:186)
> {code}
> With this patch, we don't want to copy over the table properties or any other
> physical properties (storage handler e.t.c) of the source table when CREATE
> TABLE LIKE TABLE command is executed. Only table schema (logical properties)
> should be copied. This implementation is consistent with other SQL engines
> like Mysql, Redshift etc.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
