dpQuerySplit() when using NGA
The same syntax, meaning the same dpQuerySplit() is used for NGA as for HDB/RDB. However, the following applies to the dpQuerySplit() when used with NGA.
- In NGA projects, a query is not automatically restarted after a redundancy switch.
- When using NGA direct the query a redundancy switch does not interrupt the data requests.
If data for several archive groups and/or backends is queried, the configured split size (number of values per split packet) must not be used.
The NGA Manager retrieves data for multiple archive groups/backends asynchronously.
If data is received from all groups/backends for a certain period of time, this data is sent to the process that started the query. Therefore, in this case the split of responses is made on the basis of time slices and not on the number of values (defined split size).
- There is no exact indication for the progress. It can only be 50 or 100.
- A query with element names in the SELECT statement will always receive only one response, even if a split query was issued.