How to change the data connector type on an existing dataset

Is it possible to change the data connector type on an existing dataset?  For example, currently you are using file upload to populate the dataset from an xls file.  Is there a way to keep the existing dataset, but change the connector to a dataset via email connecter?  That is, keep the same dataset, but instead of having to upload the xls file, update the data via the "DataSet via Email Connector"

 

Thank you in advance!

D. Schlatter

Comments

  • Create the email connector dataset. Then create a dataflow using Use the xls and email dataset as inputs. Union them in a transform. Finaly. move your card over to the dataflow.

  • Thank you for the reply, an interesting work around ... but i was looking to avoid creating new dataflows, moving the card, etc.

    The dataset and the card are fine, i just want to change the way the data is being updated from a file upload to a dataset via email .... 

  • Unfortuneatly that is not possible. I always suggest using a dataflow layer as a best practice, this will make backend changes much easier to manage now and in the future, as well as keep it seamless to the users. Wish there was another solution.

  • that's an interesting suggestion, thank you.  i find it more so as we have realized that the owner of a Dataset via email does NOT get notified when the email update does NOT execute for any reason.  That is, whereas the dataflow owner can get an email notification when the dataflow fails to update, this is NOT available for Datasets via email.

    Any thoughts on how to wrap a dataflow around an Dataset via email connector?  The intent being to get an email notification if for some reason the emailed dataset fails to update the dataset ... 

  • Yes. I've built this solution out for a few customers in the past. Essentialy what you need to do is leverage the domostats connector. It has a dataset run history. From that:

    1) Derive the last run for each dataset in a transform.

    2) Find the delta from last run to now.

    3) Create a card and make an alert with the rule If the time difference is greater than (your cadence i.e. 24hr) send an email to whoever you like.

     

    An less flexible but simplersolution would be:

    1) Derive the last run for each dataset in a transform and incluse the status/stage field

    2) Filter for all with status/stage != 'Successful'

    3) Create a card and make an alert with the rule to look at a count of rows > 1 then send an email.

     

    Tried to keep it high level so hopefully that makes sense.