Re-Reading large databases on startup takes a lot of time

[ohrenkino]

See here for some numbers and times:

Sluggish performance

Couldn't you supply some real benchmarks? Like: how long does it (really) take to read the file? I just tried to read some 200.000 files from an external USB3.0 HDD and it took 10 minutes, 55 seconds if MP3tag resides on the same external drive 8 minutes, 59 seonds if MP3tag resides on the internal SSD. It becomes 6 minutes 43 seconds if MP3tag resides on an internal SSD and the data also on an internal HDD. It looks to me like you can expect to read some 20.000 files/minute.

Filling the library initially takes as long as reading all the files without the library. From then onwards, reading a lot of known files should be much quicker with the library in use. New files have to be read in full - which is the slowest.
So I think that this function ...

@Johnk860

There should be a way to establish a change log that identifies changes made in Mp3tag and updates only those on startup, rather than reading the entire database as a completely new file that replaces the existing one

is already implemented.