concept field in category azure

This is an excerpt from Manning's book Azure Storage, Streaming, and Batch Analytics: A guide for data engineers.
When the structure of the data storage format changes over time, this is data drift. “Data drift exists in three forms: structural drift, semantic drift, and infrastructure drift.”[8] The set of fields contained in a data file can increase or decrease over time. This is structural drift. The content of each field can contain new values with the same meaning, or new meaning. This is semantic drift. And the systems generating, housing, or processing the data may change, leading to entirely different formats. This is infrastructure drift. Since computing systems change over time, data drift is a natural part of operating computer systems. Data drift causes problems with data analysis, because two sets of data structures representing the same type of content must be handled differently.
The segmentation of the data is driven by the distribution key. This key is one or more fields that are included in the table. In general, the key should be chosen based on the intended query usage. For lookup tables where a single row is used with a join, the unique identifier for the join is a good distribution key. For lookup tables where multiple rows are returned by the query, a field representing a foreign key is a good choice, when used with the hash distribution method. For data tables not reliant on joins, but that are used with
WHEREclauses to define ranges like “greater than” or “between,” use a field that represents a continuous set of values, like a date field or integer.