Any idea if it would be possible to use flat files as the historian in Ignition? I understand high performance historians like PI use something like that, so I read about these open-source TeaFiles (http://discretelogics.com/teafiles/) and thought if it would be possible to do the same in Ignition. Having an API available it should be possible to develop some sort of JDBC connector, but I get lost in the details or differences between SQL databases and flat files when it comes to JDBC.
Ignition’s historian is tied strictly to off-the-shelf relational databases precisely to avoid the binary lock-in of flat files
If you just want a file-based DB, you can configure a SQLite database driver & connection, and point the historian at it. You could also use something like Kymera’s InfluxDB historian module if you really need more performance - but generally, the advantages of the RDBMS approach outweight the performance downsides.
Attempting to bridge the historian into TeaFiles in particular would 100% require a custom module; you’re not going to shim the required JDBC APIs in a suitable way, so you’d have to implement your own Historian implementation via a custom module (as Kymera did) that uses Teafiles as a backing mechanism.
Yes, lucky enough to know the Kymera module very well. Derek/Kyle were extremely helpful when testing it and, performance-wise is impressive. I’m also testing TimescaleDB (it’s basically Postgre, so the JDBC driver works out of the box), and trying to get hand on GridDB AE…so it was not because of the lack of alternatives, but pretty much out of curiosity about the limits of Ignition.
Regards.
PD: Had not thought about the SQLite trick…will give it a try and see how it fares.
