5 Replies Latest reply on Jan 26, 2017 10:53 AM by cmorfin RSS
    ilopez Explorer

    Error_uploading_new_dataset_in TWA_mainMashup

    Hello,


    I am using TW Analytics Server on cloud, the trial version for 30 days.

    After following the "TW_AnalyticsBuilder_Quickstart" tutorial, I have tried to create a new dateset with two related .csv and .json files, formatting them taking as references the .csv and .json files available in the Bean_Pro_Demo tutorial.


    In the TWA_MainMashup, when I try to upload  my .csv and .json files, the platform returns the following message:


    "Error invoking service: JavaException: org.apache.http.HttpException: { "naked-array": {"errorId":"e729607c-b2ca-4fc8-80f7-7ea1d4ce5341","errorMessage":"Field name [identifier] is marked as unmodifiable and cannot be updated"} } (CreateAndConfigureDataset#54)"

     

    Could please someone help me in finding where is the error and why the platform does not accept these files?


    Thanks in advance.

    Ivan

      • Re: Error_uploading_new_dataset_in TWA_mainMashup
        ttielebein Collaborator

        Can you see if there are any more errors in the Application Logs? It is under Monitoring on the ThingWorx Composer home page

          • Re: Error_uploading_new_dataset_in TWA_mainMashup
            ilopez Explorer

            I tried again just right now and in the Application Log I can read these similar messages:

             

            at time: 15:59:33

            [context: com.thingworx.webservices.context.HttpExecutionContext@62114c46][message: Execution error in service script [TW_ML_Helper CreateAndConfigureDataset] : JavaException: org.apache.http.HttpException: { "naked-array": {"errorId":"870e102c-1eab-4aed-97af-2c9beaaf746a","errorMessage":"Field name [identifier] is marked as unmodifiable and cannot be updated"} } (CreateAndConfigureDataset#54)]

             

            at time: 15:58:55

            [context: com.thingworx.webservices.context.HttpExecutionContext@259c9096][message: Execution error in service script [TW_ML_Helper CreateAndConfigureDataset] : JavaException: org.apache.http.HttpException: { "naked-array": {"errorId":"b0ad40d3-3cfe-4e7b-b0d6-0fd77b6b431e","errorMessage":"Field name [identifier] is marked as unmodifiable and cannot be updated"} } (CreateAndConfigureDataset#54)]

             

            Thanks for your help.

            Ivan

              • Re: Error_uploading_new_dataset_in TWA_mainMashup
                cmorfin Communicator

                Hi Ivan

                 

                Note that TWA has got two special fields: identifier and entry_dt.

                Those two fields need to be present in the dataset csv file however they should not be declared in the dataset json file.

                From the error message I think you declared identifier in the json.

                You should simply remove its definition form the json.

                 

                If it does not help, is it possible that you post the json file and the headers of your csv file ?

                 

                Thanks

                Christophe

                  • Re: Error_uploading_new_dataset_in TWA_mainMashup
                    ilopez Explorer

                    Hi Christophe,

                     

                    The previous problem seems to be fixed, but now the upload of both .csv and .json files seems not working.

                    Indeed, the platform returns a "failed" message, as showed below:

                    The first example (BeanProDemo is the one provided by PTC) works greatly.

                    Right now, I am just working on some csv examples found on internet, so I can attach the csv and json file of "sacramento" example as you asked.

                     

                    Thank you for you help.

                    Ivan

                      • Re: Error_uploading_new_dataset_in TWA_mainMashup
                        cmorfin Communicator

                        Hi Ivan

                         

                        If you select the failed job and select Job Details, you will have more details on the error.

                        In most cases, this give you a clue on what is going on.

                         

                        In the case of your Sacramento dataset the issue is because the sale_date field is defined as DOUBLE in the json file but the actual values are STRING, so changing this in the json will resolve the problem.

                        Note that you also have a similar issue with latitude and longitude defined as DOUBLE while thy are not valid double values, so should also be changed to STRING.

                         

                        You should then be fine to upload this dataset.

                        Kind regards

                        Christophe