[
https://issues.apache.org/jira/browse/NIFI-7171?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17043534#comment-17043534
]
Gregory Kline edited comment on NIFI-7171 at 4/30/20, 4:29 PM:
---------------------------------------------------------------
Expanding on the behavior and expectations:
When this exception is encountered, the processor rolls back the session,
returning the incoming flow file to the queue. This behavior is ultimately
what is most problematic for me, as there is no way to provide alternate
processing for the file.
It would be acceptable, however, for the processor to transfer the flow file to
"failure" relationship with an attribute set containing the error message. In
this way, at least an alternate processing path can be provided.
was (Author: gregory.kline.arcadia):
Expanding on the behavior and expectations:
When this exception is encountered, the processor rolls back the session,
returning the incoming flow file to the queue. This behavior is ultimately
what is most problematic for me, as there is no way to provide alternate
processing for the file.
For other database types such as MySQL, the processor successfully generates a
sql statement and transfers to the success relationship, and ideally it should
behave the same for the SQL Server adapter.
It would be acceptable, however, for the processor to transfer the flow file to
"failure" relationship with an attribute set containing the error message. In
this way, at least an alternate processing path can be provided.
> GenerateTableFetch produces SQLException when table is empty
> ------------------------------------------------------------
>
> Key: NIFI-7171
> URL: https://issues.apache.org/jira/browse/NIFI-7171
> Project: Apache NiFi
> Issue Type: Bug
> Components: Extensions
> Affects Versions: 1.10.0, 1.11.2
> Reporter: Gregory Kline
> Priority: Major
>
> When GenerateTableFetch executes with the following conditions ...
> # partition size = 0
> # max value columns are unset
> # connecting to MS SQL Server
> # the target table has 0 rows
> ... the processor throws a SQLException "Unable to execute SQL select query
> SELECT -1 FROM [mydb].[dbo].[tbl_test_01] due to java.sql.SQLException: No
> rows returned from metadata query: SELECT -1 FROM [mydb].[dbo].[tbl_test_01]"
> and rolls back the flow file.
> The issue appears to be introduced in the fix for
> https://issues.apache.org/jira/browse/NIFI-5855 when -1 was used in place of
> "COUNT( * )". For an empty table, the SQL Server driver returns 0 rows for
> the query "SELECT -1 FROM <table>". GenerateTableFetch throws a SQLException
> when the resultset contains 0 rows.
>
> Version details:
> MS SQL Server: 12.0.6108.1
> JDBC Driver: com.microsoft.sqlserver:mssql-jdbc:7.4.1.jre8
> JDBC Class: com.microsoft.sqlserver.jdbc.SQLServerDriver
>
> Edit: 4/30/2020: After further testing, I have replicated this behavior
> wi{color:#172b4d}th postgres:8.4 and mysql:8.0.13{color}
> {color:#008000} {color}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
