Learn approaches to data retrieval (distribution strategies)

The interface provides several strategies for retrieving data, depending on the format of the result rows that your query returns, which in turn depends on the structure of the RDBMS tables that you are querying. The various strategies enable you to keep the impact of the interface on the RDBMS to a minimum by writing the maximum amount of data using the smallest possible number of queries.

Data retrieval strategies provided by the interface are as follows:

• If the result rows contain timestamp, value, and/or status, use single strategy.

• If result rows contain a single timestamp, tag name, and value, use distribution.

• If result rows contain multiple values and no target tag names, use tag groups.

• If result rows contain multiple values and tag names, use RxC

For distribution and RxC, you define a distributor tag, which executes a query that retrieves values to be written to target tags but does not receive any retrieved value itself. The distributor tag must have a string or numeric data type (Float16, Float32, Float64, Int16, Int32), and its timestamp is always the current time. After its query is executed and the results are distributed to target tags, the distributor tag is updated with two events indicating the results of the operation:

• The number of events that were successfully distributed to target tags

• The number of rows in the query result set

Both entries are timestamped with the current time, and appear in the archive at the same timestamp. To optimize the performance of the interface, choose an approach that retrieves the maximum amount of data using the fewest queries possible.