Good question -- no, this must be an artifact of running InferAvroSchema
before the QueryRecord processor. InferAvroSchema infers the schema and
places it in a flow file attribute, which the CSVReader uses in its
configuration.
I just debugged nifi and found that it thinks one of my fields is
Joe,
Thanks. I was going to suggest we spawn off a new thread if this continued with
more back-and-forth.
I did notice that in the JIRA you indicated that
you are using a CSVReader service to read the data. In the message, it
indicates that the FlowFIle's
filename is
I'm changing the subject name so as not to muddy the [VOTE] thread.
I finally found the error in the logs. This was also not happening prior
to the upgrade. I thinkI'll debug the server to determine the root cause:
2017-10-02 16:09:47,004 ERROR [Timer-Driven Process Thread-81]
Hey Joe,
I've seen this error crop up a time or two before. The issue is a bug in how we
are handling a failure
condition. Haven't yet figured out the best way to tackle it, but the thing to
note is that this error occurs
when there is some other sort of failure -- typically, a failure to
Apache NiFi Community,
I am pleased to announce that the 1.4.0 release of Apache NiFi passes with
10 +1 (binding) votes
9 +1 (non-binding) votes
1 0 votes
0 -1 votes
Thanks to all who helped make this release possible.
Here is the PMC vote thread:
I have to change my non-binding +1 vote to a 0, unless there is a
workaround for the bug [1] I just submitted, regarding the QueryRecord
processor being broken.
[1] https://issues.apache.org/jira/browse/NIFI-4452
On Mon, Oct 2, 2017 at 3:23 AM, Kevin Doran wrote:
>