Which is an example of schema drift?

Which is an example of schema drift?

• A. The average value of an attribute has gone up by an order of magnitude, and now your summary statistics report no longer makes sense.
• B. A data integration workflow is failing more and more due to overwhelming data volumes.
• C. Different departments store the same dataset in different formats and cannot agree on which dataset should be the source of truth.
• D. An external partner's dataset used by your organization is overhauled and many of the attributes are renamed.

Answer: (B)

Schema drift means the data you consume no longer fits the design and expectations of your pipeline, causing failures or the need to adjust the schema or logic. When a data integration workflow starts failing as data volumes grow, it reveals drift in how the data interacts with the pipeline: the system was built for a certain load and set of processing assumptions, and as the data stream accelerates, those assumptions are violated. This exposes a mismatch between how the data behaves under higher volume and what the pipeline can handle, which is a form of drift in practice. While other options show changes to the actual schema (like renamed attributes or different formats), this scenario highlights drift through changing data characteristics and the pipeline’s response to them, namely failures due to volume.

Schema drift means the data you consume no longer fits the design and expectations of your pipeline, causing failures or the need to adjust the schema or logic. When a data integration workflow starts failing as data volumes grow, it reveals drift in how the data interacts with the pipeline: the system was built for a certain load and set of processing assumptions, and as the data stream accelerates, those assumptions are violated. This exposes a mismatch between how the data behaves under higher volume and what the pipeline can handle, which is a form of drift in practice. While other options show changes to the actual schema (like renamed attributes or different formats), this scenario highlights drift through changing data characteristics and the pipeline’s response to them, namely failures due to volume.