Cannot get data from azure data lake gen2 Common Data Model

Question

Cannot get data from azure data lake gen2 Common Data Model

OUADAH Djilali 25

Hello,

I tried to export data from my data lake storage with a dataflow in Azure Data Factory, the data structure seems good and I have the lines but I don't have any data appearing in the fields:

User's image

I have tried with the scheme derivation and without it, the configuration seems fine: User's image

The only thing that bother me is that I don't have "Common Data Model" so I have selected this option but it looks really similar:

User's image

The files are automatically updated by the Microsoft Power Platform environment: User's image

Thank you for your help!

Best regards,

Djilali.

0 comments

Answer accepted by question author

1 additional answer

Your answer

Answer 1

Pilladi Padma Sai Manisha 6,430 Microsoft External Staff Moderator

Hey OUADAH Djilali,

it looks like the reason you’re getting rows full of NULLs is that you’ve hooked up the wrong inline dataset type in your Mapping Data Flow. “Common Data Service” is the Dataverse/OData connector – it won’t actually read your CDM model.json + its CSV partitions in ADLS Gen2. To pull real values from a Power Platform-generated CDM folder you need to use the Common Data Model file format.

Here’s what I’d try:

In your Data Flow, edit the Source transform and choose Inline dataset.
For Type of inline dataset, pick File, then under Format select Common Data Model.
- If you don’t see “Common Data Model” in the drop-down, choose an ADLS Gen2 inline dataset and then set Format to Common Data Model.
Point the linked service to your Data Lake Storage Gen2 account.
Under Options de la source → Format des métadonnées, pick Model.json (or try Manifest if you have a manifest file).
In Emplacement racine, browse exactly to the folder containing your model.json (not to a parent or child folder).
Expand Référence d’entité and pick your entity (crf19_registration_distributor).
Switch to the Projection tab and click Import Projection (or turn on “Infer drifted column types” if you need to tweak date/time formats).
Go to Aperçu des données and you should now see actual values in the columns.

A few additional troubleshooting tips if you still get NULLs:

• Make sure the CSV partitions in your CDM folder all have the correct number of delimiters (N–1 commas for N columns).

• If your model.json partition URIs include unsupported characters (like @snapshot=…), remove that from both the file name and model.json.

• Confirm your ADF/Synapse managed identity or service principal has at least “Storage Blob Data Reader” on the container.

Give that a shot and let me know what you see!

References

Common Data Model format in Azure Data Factory: https://docs.microsoft.com/azure/data-factory/format-common-data-model
Source transformation in mapping data flow: https://docs.microsoft.com/azure/data-factory/data-flow-source
Troubleshoot CDM “No output data” & corpus path errors: https://docs.microsoft.com/azure/data-factory/data-flow-troubleshoot-connector-format

Note: This content was drafted with the help of an AI system. Please verify the information before relying on it for decision-making.

OUADAH Djilali 25 Reputation points

2026-04-01T07:30:19.49+00:00

Hello,

Thank you for your response, I understand that I'm in the wrong inline dataset, but I tried different file options but it's not doing what I'm looking for, I only have a limited list of inline dataset options I have the impression that I am missing some of them:

Djilali,
Pilladi Padma Sai Manisha 6,430 Reputation points Microsoft External Staff Moderator

2026-04-01T22:55:48.1666667+00:00

Hi OUADAH Djilali,
Thank you for the additional details and screenshots, they help clarify the behavior you’re seeing.

Based on your latest update, the reason you don’t see the “Common Data Model” option in the inline dataset list in Azure Data Factory is expected in certain environments. This is not something you are missing in your configuration; rather, it is a known limitation where Mapping Data Flows do not always expose CDM as an inline dataset format.

Because your data is exported from Dataverse into ADLS Gen2 using the Common Data Model structure, the data itself is physically stored as partitioned CSV files referenced by the model.json. When using the “Common Data Service” connector, ADF attempts to connect to Dataverse directly, not to the files in storage, which is why you are seeing schema but all values as NULL.

To move forward, the recommended approach is to read the underlying data files directly. You can do this by selecting an inline dataset of type “DelimitedText” and pointing it to the folder that contains the CSV partitions for your entity (for example, the crf19_registration_distributor folder). Once you import the schema from those files, you should be able to preview actual data instead of NULL values.

Alternatively, if you prefer to stay closer to the CDM structure, you can use a Copy Activity to ingest the data from ADLS Gen2 and then process it downstream, as Copy Activity has more consistent support for these file-based scenarios.

you are not missing any dataset options. The behavior is due to how Mapping Data Flows currently handle CDM sources. Reading the partitioned files directly is the most reliable way to access your data in this case.

Please let me know if you need help identifying the correct folder path or validating the file structure.
Pilladi Padma Sai Manisha 6,430 Reputation points Microsoft External Staff Moderator

2026-04-02T22:31:03.05+00:00

Hi OUADAH Djilali,
Following up to see if the below answer was helpful. If this answers your query, do click Accept Answer and Yes for was this answer helpful. And, if you have any further query do let us know.
OUADAH Djilali 25 Reputation points

2026-04-07T09:45:38.1166667+00:00

Hello @Pilladi Padma Sai Manisha , the only option that seems to be available and no too difficult to put in place for the moment is to directly work with the CSV file with a copy action. I just finished to put it in place.

Thank you for your help.

Answer 2

The behavior (schema visible but all values NULL) matches a known Common Data Model issue in Azure Data Factory mapping data flows when the CDM folder only contains physical models and not a logical model.

In this case, the CDM source can read the entity definition and build the projection, but it cannot resolve the data partitions correctly, so no data rows are returned.

To fix the issue:

Ensure the CDM folder is separated into logical and physical models.
- The logical entity definition should describe the entity and attributes.
- The physical (resolved) entities and their data partitions should be described in a manifest that points to the actual CSV files.
Use a logical model as the entity reference in the data flow source:
- In Source → Source options, keep Metadata format set to Model.json (or manifest if applicable).
- Point to the manifest/model.json that contains the logical entity definition.
- Ensure that this manifest also describes the locations of the resolved entities and their data partitions.
Validate that the data partition paths in the manifest/model.json point to ADLS Gen2 (*.dfs.core.windows.net) and not to blob (*.blob.core.windows.net). Incorrect corpus paths can also prevent data from being read.

If the current export from Power Platform only produces physical models, regenerate or adjust the CDM metadata so that a logical model and corresponding manifest are present, then reconfigure the ADF source to use that logical model.

References:

OUADAH Djilali 25 Reputation points

2026-03-31T14:59:38.65+00:00

I followed the https://learn.microsoft.com/power-apps/maker/data-platform/export-to-data-lake-data-adf tutorial, the system automatically get the entity when I configure the data source but I still don't get any value in the fields..

I don't think it's an issue with the CDM as I can access it through Power BI.

Share via

Cannot get data from azure data lake gen2 Common Data Model

1 additional answer

Your answer