[
https://issues.apache.org/jira/browse/SPARK-44262?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17739511#comment-17739511
]
Florent BIVILLE commented on SPARK-44262:
-----------------------------------------
Thanks for sharing the document, much appreciated!
I know and co-maintain the Spark connector for Neo4j, which is indeed the most
straightforward way to interact with Spark.
My investigation is specifically about JDBC, which is why I have not used the
Spark connector for Neo4j here.
> JdbcUtils hardcodes some SQL statements
> ---------------------------------------
>
> Key: SPARK-44262
> URL: https://issues.apache.org/jira/browse/SPARK-44262
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 2.2.0
> Reporter: Florent BIVILLE
> Priority: Major
>
> I am currently investigating an integration with the [Neo4j JBDC
> driver|https://github.com/neo4j-contrib/neo4j-jdbc] and a Spark-based cloud
> vendor SDK.
>
> This SDK relies on Spark's {{JdbcUtils}} to run queries and insert data.
> While {{JdbcUtils}} partly delegates to
> \{{org.apache.spark.sql.jdbc.JdbcDialect}} for some queries, some others are
> hardcoded to SQL, see:
> * {{org.apache.spark.sql.execution.datasources.jdbc.JdbcUtils#dropTable}}
> *
> {{org.apache.spark.sql.execution.datasources.jdbc.JdbcUtils#getInsertStatement}}
>
> This works fine for relational databases but breaks for NOSQL stores that do
> not support SQL translation (like Neo4j).
> Is there a plan to augment the {{JdbcDialect}} surface so that it is also
> responsible for these currently-hardcoded queries?
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
