I would like to have a tool for collection database clean-up.

I have a relatively large colection with 17 000+ songs. Database file is 14,5 MB size and amaroK after 0,5 years of database creation got very slow. There is a 0,5 - 1 sec delay on every search. Half a year ago amaroK was instant!!
I removed database file and created new collection. Database file was 12 MB size now and amaroK was realy faster. Apparently I have lost all my statistics, ratings and scores.

I restored old database file and check it with raw hex viewer. And I found TONS of records of songs and albums I have deleted ages ago!! I did rescan-collection many times, but it seems that amaroK doesn't remove records about non-existant files from database. I think it would be useful to have some tool for database cleaning up and riding off records about non existant files. amaroK would be really faster for me and for all of us. That 0,5 sec delay on every search really annoys me.

I agree, this would be a welcome addition, no doubt there's something you could do on the actual database but to make it easy, why not put it in an admin settings section.  While we're at it, it would be great to have a backup option for those statistics perhaps with selectable backup options such as covers, stats etc.  Those two items together would be awesome!  No more losing stats, data, rescanning covers etc!  Yeeha!

You did not mention what kind of database backend you are using, but if it's SQLite, you should definitely switch to MySQL or Postgres. SQLite does not currently scale very well with big collections.

How do I use MySQL?
Do I have to set up a MySQL-Server?
How do I do that?
Should I use MySQL with about 4000 Songs or or should I stay with SQLite?

deloco wrote:Should I use MySQL with about 4000 Songs or or should I stay with SQLite?

4000 tracks is not so much. SQLite should be fine.

So at what Track-count you recommend using MySQL?
When I add my Audiobooks to my Collection it's about 5000 Tracks, I believe it's not so much either is it?

That's a good point actually, it would be nice if there was a track count recommendation on the initial setup screen so we knew when to use each type of database.

It's hard to say. Rule of thumb: If your collection is bigger than 10000 tracks, you might consider trying MySQL or PostgreSQL.

Okay... Till I got 10000 Tracks it will take a while... So I'll stay with SQLite... Thanks so far.

And what about the difference between mysql and postgresql with ~10000 tracks ? Any idea ?

nonor wrote:And what about the difference between mysql and postgresql with ~10000 tracks ? Any idea ?

Nope, no idea. Be the first to find out


(PostgreSQL is often said to be the better database, though)

Does MySQL or Postgresql solve the problem of non existant files being kept in database ?