Metrici's internal structures are based on a triple store. This is very flexible, but makes it difficult to visualise and process data aggregated from multiple nodes. To make it easier, Metrici provides conventions and utilities for processing and visualising aggregated data as tables. These utilities are grouped into the Table (library.table).
Metrici tables are very versatile, and can be used to represent tabular data, extract data from nodes, perform set-based calculations, and display data as tables, in summary form, or as charts.
- Table structure
- Table nodes
- Table views and columns
- Table member types
- Table display options
- Programming with tables
Metrici's support for tabular structures has evolved over the years. Old version of components and current but deprecated versions use different conventions.
Specifically:
- Parts (library.parts) provides support for an earlier version of column scripts that were used solely for display purposes. These are used by some now-deprecated package types which have been superceded by Package (library.core.PackageType).
- Result set (library.resultset) provides support for JSON result set structures for managing and displaying result sets. These are being replaced by tables.
- Assessment supports analysis items and analysis scripts, which are similar in intent to table views and column scripts used to extract data. The newer table components offer order-of-magnitude performance improvements.
- Table supports an older convention for building tables based on one-row tables known as "tuples". This has been superceded by column scripts, though existing tuples can continue to be used either as one-row dynamic tables or to specify columns within a view.
Use the components documented in this section for new solutions. When examining old solutions, watch out for the older components above.