hash function key in cached files

[misamae]

I've copied an old folder which contains cache files generated by memoise to a new system. For some reason while memoise generates different cache keys that does not match existing cache key. Can you give a bit of insight on how the cache keys are generated?

[wlandau-lilly]

The cache keys are generated by hashing, as implemented in the digest package. A memoized function computes a hash from the original function body, the function arguments, and user-defined formulas that tweak caching behavior. This hash is the cache key, which is also the file name of the cached object in the file system cache. Some platforms treat file names differently, but the characters in hashes should theoretically be safe. Would you walk me through the specifics of your use case? A minimal working example would help if possible.

[misamae]

Thanks for the explanation. It didnt occur to me that function body was part of the hash. In my head it was only function arguments and I was wondering why the code doesnt hit the old cache. We have had change the function body few times. However, I believe the changes were only new lines addition and removal.

[jimhester]

If the function changes the cache should be invalidated, so this is working as intended.