Hi, We have event-based data currently being logged in relational database structures. For various reasons this data need to be stored elsewhere, possibly on our existing PI historian database. We need to ascertain whether PI can relatively easily store this type of event based data, and ultimately produce reports using RT Web Parts, also relatively easily. If not, we would rather log it in a SQL Server database. 1. This data are obtained via a SSIS package from flat files and stored in a SQL server database. 2. Currently we have around 550 event based jobs running parallel typically on a daily basis. 3. This constitutes in approximately 100 000 records every 5 minutes being logged. That’s more or less 350 records every single second. 4. These events are currently sorted on a BigInt type incremented number from low to high. Can PI log values at a micro sub second level so that the values can be displayed in the correct order it was logged? Or are there some other way to do this? 5. Note that EvTime only saves at second level. EvSeqNo is necessary here to sort the data sequentially, as many records are logged per second. 6. Can Pi handle this huge influx of data? 7. A sequence can constitute a various number of events, from starting that sequence by starting a job by an operator, holding or diverting it. Starting, holding, diverting and so forth constitutes an event. There are various other events as well 8. A job can start another job as well. 9. EvJob, EvLevel, EvLabel, EvKey,EvText, EvSeqNo are all stored in separate database columns. Must this be concatenated into one delimited string or is there an easier way to store the data in PI? 10. EVJob should typically the tag name. However when a job is started by an operator EVJOb can be something like :Citect which means that the actual job being started must be retrieved from the EvText string. 11. Once all this data is being logged, can it easily be retrieved to display in reports using RT Web Parts?
