PI Vision currently offers 3 table symbols:
Table: tags or element-attribute pairs along the rows, properties and statistics along the columns
Asset comparison table: elements along the rows, attributes along the columns
Time series table: 1 event per row of a single tag or a single attribute of a single element
The "table" symbol came first, and the other 2 came later, with each one offering a different way to display data in a table. Each new type of table is an incremental improvement to PI Vision. However, it is confusing to keep track of what each table is for, especially the original "table" since its name does not have any further description. Also, on the feedback website, there are many requests for improvements to the table symbols or for additional table symbols. It is infeasible for AVEVA to implement these as incremental improvements. It would be too much patch work and too confusing for users.
Instead, AVEVA should replace these table symbols with a single generalized table symbol that is customizable enough to create basically any shape/configuration of table that the user might desire.
Use Excel PivotTables as an example. They let you put any field along the rows, columns, and filters and let you specify the type of aggregation to perform for each of the fields in the PivotTable's cells. There can be multiple fields along the rows and columns, creating an expandable/collapsible hierarchy that can optionally have subtotals and grand totals (or, more generally, an aggregation at the parent level). They also let you sort and filter by any value of any field that is used in the PivotTable. The concept is simple, general, and powerful.
However, the generalized table symbol in PI Vision cannot copy Excel PivotTables exactly since there are considerations that are specific to PI, such as:
Aggregations would need to include not just the traditional sum, average, count, minimum, maximum, standard deviation, range, etc. but also things like the value at the start of the time range or the value at the end of a time range.
The different data retrieval methods for time-series data would need to be considered if the data is not being aggregated (e.g. the raw archived events, sampled values at regular intervals, archived data at or before/after a list of regularly-spaced timestamps, etc.).
Basically, much of the functionality of PI DataLink would need to be replicated in PI Vision's generalized table symbol, which only makes sense since PI DataLink is for Excel, and Excel is all about tables.
Lastly, the install kits of versions of PI Vision that have this generalized table symbol should convert the existing table symbols, in any visuals that use them, to the generalized table symbol. The generalized table symbol should be called "Table" or "Pivot Table".
I realize that implementing a generalized table symbol would be a lot of work, but the end result should be worth it!
