You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Stephen Durfey (Jira)" <ji...@apache.org> on 2022/08/09 20:33:00 UTC
[jira] [Created] (SPARK-40024) PostgresDialect Doesn't handle arrays of custom data types after postgresql driver version 42.2.22
Stephen Durfey created SPARK-40024:
--------------------------------------
Summary: PostgresDialect Doesn't handle arrays of custom data types after postgresql driver version 42.2.22
Key: SPARK-40024
URL: https://issues.apache.org/jira/browse/SPARK-40024
Project: Spark
Issue Type: Task
Components: SQL
Affects Versions: 3.1.1
Reporter: Stephen Durfey
Starting in version 42.2.23 (also 42.3.x and 42.4.x), the sql type returned by the postgresql driver is now `ARRAY` with columns with an array of custom data types (e.g. an array of enums). Prior to this version the driver returned the type `CUSTOM`. PostgresDialect can handle custom types and array types, but not array of custom types. Tthe type support within arrays is limited to what is listed in the catalyst type here: [https://github.com/apache/spark/blob/master/sql/core/src/main/scala/org/apache/spark/sql/jdbc/PostgresDialect.scala#L69-L98.] Since a custom type won't match any of those, `None` is returned and eventually `JdbcUtils` throws this exception:
```
java.sql.SQLException: Unsupported type ARRAY
at org.apache.spark.sql.errors.QueryExecutionErrors$.unsupportedJdbcTypeError(QueryExecutionErrors.scala:682)
at org.apache.spark.sql.execution.datasources.jdbc.JdbcUtils$.getCatalystType(JdbcUtils.scala:249)
at org.apache.spark.sql.execution.datasources.jdbc.JdbcUtils$.$anonfun$getSchema$1(JdbcUtils.scala:327)
at scala.Option.getOrElse(Option.scala:189)
at org.apache.spark.sql.execution.datasources.jdbc.JdbcUtils$.getSchema(JdbcUtils.scala:327)
```
The postgresql driver change was part of [Issue#1948|[https://github.com/pgjdbc/pgjdbc/issues/1948].]
I did make a change locally and returned `StringType` instead of `None` for the default case, and that worked fine, but I don't know if that's the desired solution or not.
I created a gist with a code snippet to recreate the issue and run it via spark-shell: https://gist.github.com/sdurfey/f9e73cffaeb90cd9c69dcc771fe59f08
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org