I wrote up an Improvement Jira to add the property to Validate Field
Names to AvroSchemaRegistry:
https://issues.apache.org/jira/browse/NIFI-4612

-Matt

On Thu, Nov 16, 2017 at 9:15 AM, Matt Burgess <[email protected]> wrote:
> Mike,
>
> That's a very good point, I guess that would have to be emulated
> somehow. I wonder how Linkedin and Coursera (who use a Scala binding
> for Pegasus called Courier) handle all that?
>
> Anyway as it turns out Schema.Parser() has a method called
> setValidate(boolean) [1] which can enable/disable name validation.
> Perhaps we should disable it or add a property to AvroSchemaRegistry
> to allow it to be disabled?
>
> Regards,
> Matt
>
> [1] 
> https://avro.apache.org/docs/1.8.2/api/java/org/apache/avro/Schema.Parser.html#setValidate(boolean)
>
> On Wed, Nov 15, 2017 at 5:29 PM, Mike Thomsen <[email protected]> wrote:
>> Might want to be careful about that because Avro 1.8 added the support for
>> logical types and removing that could break parts of the Record API like the
>> date/timestamp functionality.
>>
>> On Wed, Nov 15, 2017 at 3:41 PM, Matt Burgess <[email protected]> wrote:
>>>
>>> Thomas,
>>>
>>> You can file an Improvement or New Feature Jira [1] asking for the
>>> enhancement. Ironically Avro 1.4.0's Schema.parse() method does allow
>>> the dollar sign, but we use Avro 1.8.x now which is apparently more
>>> strict.  I am toying around with a PegasusSchemaRegistry using the
>>> PegasusSchemaParser from Linkedin's rest.li project [2]. It is very
>>> much like Avro but has some nice differences [3] and the parser
>>> doesn't choke on invalid Avro identifiers.  Another alternative (in a
>>> separate NAR) is to provide a LessStrictAvroSchemaRegistry that is
>>> exactly like AvroSchemaRegistry but bundles Avro 1.4.0 instead of
>>> 1.8.2 (and uses Schema.parse() rather than new
>>> Schema.Parser().parse()).
>>>
>>> Regards,
>>> Matt
>>>
>>> [1] https://issues.apache.org/jira/projects/NIFI
>>> [2] https://github.com/linkedin/rest.li
>>> [3]
>>> https://github.com/linkedin/rest.li/wiki/DATA-Data-Schema-and-Templates
>>>
>>> On Wed, Nov 15, 2017 at 3:32 PM, DENIMAL Thomas
>>> <[email protected]> wrote:
>>> > Hello Mike,
>>> >
>>> >
>>> >
>>> > Thanks for your answer.
>>> >
>>> > Do you know how can i ask for an enhancement request for this feature?
>>> >
>>> >
>>> >
>>> >
>>> >
>>> > Regards
>>> >
>>> > Thomas DENIMAL
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> > De : Mike Thomsen [mailto:[email protected]]
>>> > Envoyé : mercredi 15 novembre 2017 19:58
>>> > À : [email protected]
>>> > Objet : Re: NIFI 1.4.0 - PutMongo - How to use composite key for "Update
>>> > Query key" parameter ?
>>> >
>>> >
>>> >
>>> > Also, I'll use this thread to bring something about PutMongoRecord into
>>> > the
>>> > mailing list for other Mongo users to see.
>>> >
>>> > PutMongoRecord cannot support more than document replacement updates
>>> > because
>>> > the Mongo update operators use "$" which is an illegal starting
>>> > character in
>>> > Avro. So you cannot define a Record API-compliant schema that
>>> > corresponds to
>>> > a partial update operation. You also, unfortunately, cannot do that with
>>> > extended JSON features like date handling. You have to use the Avro and
>>> > Record API-based methods for representing a Mongo date field when
>>> > handling
>>> > records.
>>> >
>>> >
>>> >
>>> > On Wed, Nov 15, 2017 at 1:32 PM, DENIMAL Thomas
>>> > <[email protected]>
>>> > wrote:
>>> >
>>> > No one with the same issue ?
>>> >
>>> > On 2017-11-04 14:18, DENIMAL Thomas <[email protected]> wrote:
>>> >> Hello ,>
>>> >> I'm using NIFI 1.4.0 with Mongodb 3.2.17.>
>>> >>
>>> >>
>>> >>
>>> >> Is it possible to use a composite key for "Update Query key"  parameter
>>> >> with PutMongo processor ?>
>>> >> My collection "primary key" is a composite key of 2 columns, SIREN &
>>> >> NIC,
>>> >> (I'm using a composite unique index to ensure uniqueness)>
>>> >>
>>> >> I want to be able to make upsert operations on that collection with
>>> >> Nifi.>
>>> >>
>>> >> Thanks for your help>
>>> >> Regards>
>>> >> Thomas DENIMAL>
>>> >>
>>> >
>>> >
>>
>>

Reply via email to