Define Semantics JTdCJTIydGl0bGUlMjIlM0ElMjJEZWZpbmUlMjBTZW1hbnRpY3MlMjAlMjIlN0Q=

In this process, you add dimensions and measure and types of pre-calculated aggregations.

Parameter/Field	Description and Topic Reference
Dimensions	A dimension is a structure that categorizes facts and measures allowing users to answer business questions. The elements of a dimension can be organized as a hierarchy - a set of parent-child relationships, where a parent member summarizes its children. You can add, name, and rename dimensions, define a hierarchy level, add levels, and set up attributes. See Adding dimensions for more details.
Attributes	You can use dimension attributes as part of your analysis to tease out more information. See Adding attributes for more details.
Measures	Measures represent values that are counted, summarized, or aggregated, such as costs. Aggregating measures allow you to calculate the Sum, Minimum, Maximum, Count, Mean, or Distinct Count for the field data you are aggregating. See Adding measures for more details
Calculated Measures	It is like any existing attribute or user hierarchy member. You can add calculated members to a hierarchy using hierarchy or attribute properties. See Adding calculated measures for more details
Hierarchies and levels	You can add calculated members to a hierarchy using hierarchy or attribute properties. Hierarchies are a useful option for reducing the complexity between attributes and helping you with drill-down behavior. See Defining hierarchies and levels for more details

After you have designed the cube, a physical design is displayed in the Refine tab.
See Physical view for more details.

Semantic Model Design Page

Refine ModelJTdCJTIydGl0bGUlMjIlM0ElMjJSZWZpbmUlMjBNb2RlbCUyMiU3RA==

In this process, you refine the semantic model by defining the partition, aggregation, and auto cache strategies and setting advanced properties.

Parameter/Field	Description and Topic Reference
Physical view	It is a separate view of a semantic model that lets you make changes over time to tune the performance of a semantic model without losing the business design or logic design of the cube. As your business generates new data, you can add data to an already-built semantic model with an incremental build. You can also create a sliding window and specify a range of data that you want to build into the semantic model. You can delete semantic models at any time; however, if there are reports built using the semantic model, they will no longer function, and you will receive a warning before you proceed. There are two ways to view and work with a cube: Logical view (default view) Physical view You can maintain the logical view of a semantic model to support business requirements and tune the physical view to support optimization requirements. The logical view shows all calculated measures and members. See Physical view for more details.
Partition Strategy	It is a flexible means of managing cubes, especially large cubes. It improves semantic model scalability by keeping the data structure small. You can create, replace or drop partition. See Partition strategy for more details.
Aggregation Strategy	Aggregates are precalculated summaries that improve query response time by preparing some answers before questions are asked. Kyvos uses query history and builds over time to create an aggregation strategy. You get system recommendations (see Semantic Model recommendations )for a semantic model based on query pattern and data pattern analysis. See Using aggregates for more details.
Cache Strategy	The cache for a semantic model is gradually populated when a semantic model is browsed. Kyvos allows you to automate the process of cache population of semantic models to reduce the query response time. In the Auto Cache Strategy section of the Refine page, you can specify Cache rules for: A single cube Historical queries Types of queries to cache Additionally, you can specify a cache build schedule for specific cubes. See Setting up and using cache rules for more details.
Properties	Properties are organized into groups. Errors can be easily identified and visible and you can click the group to see the details. Kyvos specifies three modes for properties: Default Inherited Modified See Using cube properties for more details.

Refine model page

Process Model JTdCJTIydGl0bGUlMjIlM0ElMjJQcm9jZXNzJTIwTW9kZWwlMjAlMjIlN0Q=

After refining the cube, you can build a cube. Kyvos allows you to specify the type of build for the cube:

Full data process
Incremental data process
Test data process
Update aggregates

To schedule semantic model processes, see Scheduling semantic model processes for more details.

To see different semantic model process types, see Semantic model process types for more details.

Kyvos supports building semantic models on both on-premise (Hadoop) and Cloud (Amazon, Azure, and GCP) platforms. You can also build semantic models on a Snowflake, Redshift, or BigQuery source.

Parameter/Field	Description and Topic Reference
Full data process	Creates a full, incremental, or test data process and update aggregates.
Incremental data process	Updates only the latest changes to an existing semantic model.
Test data process	Builds the semantic model on the sample data.
Update aggregates	Updates only the aggregates in a semantic model to improve query time.
other process jobs	Specify other build jobs such as Drop Partition, Cache Build or Data Profile
Job Summary	After the process job is completed, the status is displayed as Success or Failed on the semantic model designer and the Job summary .

Semantic model process page