I’m indexing about 17,000 files … I have issues running out of memory (which I keep changing the parameters for in the .app. The full autocorrect process takes ages… it would be nice if you could save the retrieved music ids / changes etc in a database first before you write and change the files… then if something bad happens you just have to reload the dbase and save again rather than running through the whole autocorrect process once more…
just a thought 
thanks.
Acoustic Ids are saved as you go along anyway so these songs will get skipped over this part at least, but other changes aren’t . Yes I can see the worth in your idea, but perhaps in the short term you should get the memory settings sorted out or process smaller batches
perhaps i’m mistaking what the OP is requesting, but isn’t this functionality already included? you just export the metadata to an XML file; reload the same files if the computer crashes or something & apply the saved XML file to see the unsaved changes reapplied.
moreover, if you’re editing data by hand [not calling up the databases or autocorrect or anything] it’s way easier to just edit the metadata in the exported XML file in excel or numbers & then reload that XML file into jaikoz once you’re done editing the metadata anyway.
also, if you’re having memory issues, maybe you should:
STEP 1
� uncheck “generate acoustic ID for songs with no matches” or whatever in the Preferences
� hide all the songs that now have a PUID [it’s a menu item…i don’t have jaikoz open at the moment, but it’s a command in a submenu, something like “hide all songs with acoustic ID”];
� run Retrieve Acoustic ID on all yr files
� run Retrieve Acoustic ID again—sometimes it’ll work the second time through, if you’re working with a large number of files, for some reason—
� show all songs.
� save, quit.
� restart yr computer so you can boot up jaikoz with no inactive memory slowing things up
STEP 2
� open jaikoz, reload folder of songs you’re working through;
� hide all the songs that now have a PUID;
� open Preferences & check the “generate ID” thing that you unchecked last time;
� run Retrieve Acoustic ID on all yr files;
� save, quit.
STEP 3
� after 24 hours have passed—that’s how long it should take for newly generated PUIDs to show up in the online database jaikoz queries for retrieval—reopen files in jaikoz;
� uncheck the “generate ID” thing in Preferences
� hide all songs with an acoustic ID;
� run Retrieve Acoustic ID;
� run it again for any files that still don’t have a match;
� show all files; save, quit.
STEP 4, if you still had PUIDless songs after step 3
� check the “generate ID” thing in Preferences
� hide all songs with an acoustic ID;
� run Retrieve Acoustic ID;
� show all files; save, quit.
� repeat step 3…remember to wait 24 hours.
[i have a library of 100,000+ songs…that’s how i got them all PUID’d without 45867 GBs of memory. restarting yr computer whenever you complete part of the task is useful. if you have the console log open in the Details pane when you run the Retrieve Acoustic ID command, you can watch which tracks it’s processing as it goes through them, which is more interesting to watch than the little progress bar.]
Good idea with the spreadsheet, and good to know that somebody is using the spreadsheet mechanism successfully