[
https://issues.apache.org/jira/browse/PHOENIX-6783?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17877778#comment-17877778
]
ASF GitHub Bot commented on PHOENIX-6783:
-----------------------------------------
rejeb commented on code in PR #139:
URL:
https://github.com/apache/phoenix-connectors/pull/139#discussion_r1736699009
##########
phoenix5-spark/src/main/java/org/apache/phoenix/spark/datasource/v2/PhoenixDataSource.java:
##########
@@ -33,21 +37,27 @@
import org.apache.spark.sql.sources.v2.reader.DataSourceReader;
import org.apache.spark.sql.sources.v2.writer.DataSourceWriter;
import org.apache.spark.sql.types.StructType;
+import scala.collection.immutable.Map;
import static org.apache.phoenix.util.PhoenixRuntime.JDBC_PROTOCOL;
import static org.apache.phoenix.util.PhoenixRuntime.JDBC_PROTOCOL_SEPARATOR;
/**
* Implements the DataSourceV2 api to read and write from Phoenix tables
*/
-public class PhoenixDataSource implements DataSourceV2, ReadSupport,
WriteSupport, DataSourceRegister {
+public class PhoenixDataSource implements DataSourceV2, ReadSupport,
WriteSupport, DataSourceRegister, RelationProvider {
private static final Logger logger =
LoggerFactory.getLogger(PhoenixDataSource.class);
public static final String SKIP_NORMALIZING_IDENTIFIER =
"skipNormalizingIdentifier";
@Deprecated
public static final String ZOOKEEPER_URL = "zkUrl";
public static final String JDBC_URL = "jdbcUrl";
public static final String PHOENIX_CONFIGS = "phoenixconfigs";
+ public static final String TABLE = "table";
+ public static final String DATE_AS_TIME_STAMP = "dateAsTimestamp";
+ public static final String DO_NOT_MAP_COLUMN_FAMILY =
"doNotMapColumnFamily";
+ public static final String TENANT_ID = "tenantId";
Review Comment:
I think PhoenixConfigurationUtil.MAPREDUCE_TENANT_ID have another usage and
can't be used here.
The TENANT_ID property is only used at spark options level. For setting the
value for phoenix connection property I use PhoenixRuntime.TENANT_ID_ATTRIB.
I prefer keep them separate and duplicate value because
PhoenixRuntime.TENANT_ID_ATTRIB come from a phoenix dependency so if its value
changes it will have an effect on spark connector.
> Phoenix5-Spark3 Connector Spark SQL Support
> -------------------------------------------
>
> Key: PHOENIX-6783
> URL: https://issues.apache.org/jira/browse/PHOENIX-6783
> Project: Phoenix
> Issue Type: Bug
> Components: connectors, spark-connector
> Environment: # Pure Spark 3.0.3
> # phoenix5-spark3-shaded-6.0.0-SNAPSHOT.jar located in $SPARK_HOME/jars
> # SQL used to create Spark Phoenix data source table is:
> {{create table Table1 (ID long, COL1 string) using phoenix options (table
> 'Table1', zkUrl '192.168.0.103:2181', primary 'ID')}}
> Reporter: yj
> Assignee: rejeb ben rejeb
> Priority: Major
>
> While conducting the Spark 3.0 Integration test using *Phoenix5-spark3-shaded
> module of project
> [Phoenix-Connectors|https://github.com/apache/phoenix-connectors]* the
> following questions occurred.
>
> The test process is as follows:
> 1. Phoenix created a table called "TABLE1" and then loaded the data into that
> table.
> 2. After that, I tried to select the data of the table using Spark.
>
> As written in the README of Phoenix5-spark3 module, the part that loads
> Phoenix data into the Spark DataFrame using the Phoenix data source seems to
> work well.
> However, when I *create Phoenix Spark data source table through Spark SQL*
> and try to select that data source table, the following error occurs:
> {{java.util.concurrent.ExecutionException:
> org.apache.spark.sql.AnalysisException: phoenix is not a valid Spark SQL Data
> Source}}
>
> I wonder if the Phoenix-Connectors does not support creating Phoenix Spark
> data source tables, or if there is any other reason for this error.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
