[Question] What is the proper way to load large amount of data from database about differencekit HOT 4 CLOSED

@shuoli84 If the diffing does not take that long (from the reference point of a user, approx. less than 500ms), you can perform the diffing on the background thread. Anything larger, DifferenceKit or any other diffing approach might not help you. Back to calling performBatchUpdates!

Also, if you don't want all the dataset in memory, again DifferenceKit might not be what you're looking for. Diffing requires that all the elements be compared and then a determination is reached on the outcome of the differences. Unless you have someway of performing comparisons (==) without hitting anything that is not in memory, you might be fine. The same goes for Hashes. If you are hitting the storage device or the network to fetch and compare the results, it'll be too slow to be of any advantage.

from differencekit.

@Sameesunkaria Thanks for the reply. I'm thinking of a solution kinda like:

• Section info can be maintained in db, and small enough to load into memory.
• Each section has a version to indicate whether modified.
• When data changed, changeset can be computed from section info stored in db, and when load each item, a db lookup or just read cache.

from differencekit.

@shuoli84 That should work. The comparisons can be made on the version number/identifier. And for the difference identifier, a primary key can also be stored in the section info.

Just to be clear, if the number of entries are too large, the diffing may still take enough time that it can cause a stutter if being computed on the main thread.

from differencekit.

@shuoli84
Did you solve the problem?
Feel free to reopen this issue, if you didn't solved.

from differencekit.