This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(176, 12%, 27%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EMC%3C/text%3E%3C/svg%3E)
I built a model using auto ML and deployed it as a batch endpoint.
When I re-use the same dataset I used to train the data I get an error:
ValueError: Length mismatch: Expected axis has 1 elements, new values have 7 elements
I assume, I have to convert the whole thing into some kind of batch first?
I didn't find documentation on how to convert a structured table dataset to a batch. Does anyone know how to do this?
Thank you for the help, best, Max
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(214.4, 55.00000000000001%, 30%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ESK%3C/text%3E%3C/svg%3E)
Hi Max, I have recently started working on Azure-ML. Is there any chance that you could post some code to see the error?...so that I will try from my end.
I think you might have dimension mismatch in the "preprocessing data step"?...or some reshaping the data issue?
@Ramr-msft
Thank you for the help. I read the resources yousent already, but they all seem to contain unstructured data, not tables ...
According to your sample I wrapped the rows with the "data" tag. I still get the same error.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(22.400000000000002, 11%, 12%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ERM%3C/text%3E%3C/svg%3E)
@Max CH Thanks for the question. Here is the document for Azure ML Batch endpoint for Batch scoring and Troubleshooting batch endpoints.
Here is a sample that does AutoML Batch Scoring that can help.
Hi Max, I don't how you shaped the data, but I tried to do the below:
Is your data somewhere expecting to be reshaped into minibatch_size=(350,1) in the second axis?...because the error says "ValueError: Length mismatch: Expected axis has 1 elements, new values have 7 elements"
I really couldn't find any documentation regarding the shape of the input data, that's why I opened this thread.
Not quite sure what you mean, do you have a sample, so I can try it?
I thought I could take a look at the error if you had some dummy data and test code(maybe you can post over Stackoverflow).
Also, the documentation regarding the shape of input data might not be present in Azure docs because the error is generic not Azure-ML specific.
The error looks more of a kind that happens during data processing. Did you try running the whole ML experiment outside of Azure cloud on small sample of data and not using autoML?...if yes, are you getting the same error?