I have a huge collection of 500 000 cards, I’ve noticed that size of collection file of it is quite big ( almost 500 MB) so today I decided to remove some decks from it to see how much size of collection file is going to be reduced, I was removing decks one by one each time rebuilding collection in the end my collection is 300MB down from 500MB but it has only 10 cards, I removed even the last deck so anki showed default empty deck but is still 300 MB in size. I opened and closed anki many times ,rebuilded, and checked for media and empty cards… and still is 300 MB. I uploaded this 10 card deck on google drive https://drive.google.com/drive/folders/1UHfydDdQjBkjw9t9CpZMpjweqTtxzeDJ?usp=sharing
I did all of it in anki desktop 2.55 RC1 on windows
Couldn’t open the package either, even with support for legacy import/export enabled
Here some info to try to delete these files (corrupted maybe?) Managing Files - Anki Manual (it might be add-on conflict in my end)
Tools>Check Database will free up space after you delete content. If you the collection is still huge after that, it means you have a large amount of reviews logged which should be visible in the stats.
I opened it in anki 2.1.43 and I did tools>Check Database and still is 300MB, it seems like anki never remove some data and .collection keeps growing and growing endlessly. So now I have 10 cards in 300 MB file
Actually, I decided to do the experiment and I exported all my decks then imported on new anki installation in hope it will get smaller… but it got bigger by 100 MB. There is obvious that there is a bug in Anki as old as anki itself and there are no tools in anki to remove and optimize redundant data. So after years of using large decks the problem is getting more apparent, with small deck, few MB in size no one notice it, 2MB or 4 MB who cares… but 300MB or 600 MB makes difference. If all this extra data is schedule information, anki should automaticity remove it, for example last 5 or 7 be only kept
It’s not a bug. When you delete cards, their history is kept around so that graphs still reflect the work you put in. If you export the deck as tatsumoto suggested, you’ll see the resulting file is only 50kb as it does not include the history.
Exactly, what I do is putting all “ghost reviews” (reviews from deleted cards) to a dedicated profile to save some space. In my case that is 1.2 m reviews. Even tough just my revlog is well above the 250mb limit. Roughly estimated this limit equals to 3.6m reviews. So just the repetition history weights more than the actual flashcard content.
I created multiple profiles just like you… I wish there were an option to remove info regarding past reviews on to keep only three last instead of 20 years of history which is not take into account as far as I know to schedule future reviews