biomass() and standStruct() fail when using multi-state, remote FIADB object with specific evalids

[hunter-stanke]

biomass() and standStruct() fail when using multi-state, remote FIADB object with specific evalids

[hunter-stanke]

Sorry for the delayed reply, and thanks for sending this my way! I'll dig into it today, and hopefully have a fix up in the next few hours.

[hunter-stanke]

I've found the issue, and just pushed up a fix.

The problem is that the list of EVALIDs ee pertains only to Alabama, i.e., none of those listed are affiliated with the Georgia inventory. When we're working with an in-memory database, this is fine, as we're simply asking clipFIA to drop all Georgia inventories and retain the selected Alabama inventories. However, rFIA's out-of-memory methods will read in states one at a time, evaluate the call to clipFIA, and then proceed with estimation. So when Georgia gets read in, we're asking it to drop all of it's inventories, leaving us with no data and hence resulting in a less than informative error message.

For in-memory databases, I don't mind the current behavior (i.e., the ability to drop selected states entirely should you choose to). Thus I've have only added a warning for these cases - the message will simply tell you what states are missing from the list of EVALIDs.

For remote databases however, I don't see much benefit to allowing states to be dropped by clipFIA, as this could be accomplished more efficiently via the states argument with readFIA. Hence, an error will be thrown in these cases. The message will tell you what states are missing, and to either add EVALIDs for the missing states or drop these states from the call to readFIA.

Could you re-install from GitHub and confirm that behavior on your end? I'm open to suggestions on clipFIA's behavior here too. If consistency between in-memory and out-of-memory approaches is more desirable, we can certainly go that route.