Generally speaking, it's better to ensure that all of your files have the same format, and only source a dataset from one folder in your cloud storage location.Different datasets should be sourced from different folders, but each folder should be homogenous in terms of its contents.Then whenever you...

Identities must be unique to a single profile. When you have Field1 and Field2, together do they always uniquely identity a single profile? (I assume that they do. If so then you should concatenate them into a single field marked as an identity either upstream of AEP or during the Data Prep phase of...

How will you prevent your arrays and the resulting profile records from growing too large? There are multiple limits applicable here: https://experienceleague.adobe.com/en/docs/experience-platform/profile/guardrails#data-model-limits Would you be better served with custom aggregates instead, e.g. fi...

Looking at the original image, it appears that your processing throughput increased on September 13th and since then has been nearly matching your actual ingestion rate? If true and that's still not sufficient bandwidth then I'd explore this method of reducing ingestion into Profile Service:https://...

I typically combine into one dataset when:the data is of the same class: profile, event, or customthe data will have the same primary identitythe data is sourced from the same systemthe data will be sourced the same way: streaming vs batch If you combine profile-class data into one dataset when all ...

The only relevant merge policies when defining audiences are the ones for "_xdm.context.profile". Of those, only one can be marked as the default.

If you look on the Datasets page for your dataset at the bottom you'll see Dataset Run IDs and their corresponding Batch IDs. For batch source connectors these are 1:1.

With additional enhancements to AEP and related apps there are fewer and fewer reasons to have multiple merge policies. From what I've seen, most combinations of use cases function perfectly fine with a single merge policy that is used for all segmentation and all integrations. Having unnecessary ad...

As mentioned above, streaming data is batched before it's written to the data lake. But it is ingested immediately into Profile and Identity Services, so there's not much reason to worry about the intervals between batches in the data lake. Nothing should be trying to read from those datasets in the...

My preference is to avoid zero length strings at all times. Otherwise this complications the logic required when testing in an audience. This is why best practice is to set a min=1 length on string fields even when optional. If you need to nullify one or fields then I'd recommend sending a complete ...