3.1.5 too slow at Select by Location

[fernacolo]

I installed 3.1.5 and tested against a very complex source tree, with more than 200k files. I added to Scan Location an old copy of this source tree, so it found around 110k duplicates. I wanted to remove all duplicates from the old copy, so I used select by location. It's taking 4h since I've started, and it's consuming only around 15% of CPU. It's using 440mb of private memory, and it performs about 80kb of I/O operations every second (zero page faults). I don't understand why Select by Location would use so little CPU and why is it performing I/O at all. I know that Select by Location works because it took about 30min in a smaller dataset, but this time I'm almost about to kill the process.

[sharoni]

Same here, waiting forever on "Select by Location" where all I see is the hourglass, already 3-4 hours.
Only 1-2% CPU and lots of I/O
This must be a bug in implementation, if not in design
bye bye Duplicate Cleaner

[DigitalVolcano]

Fixing the Select by Location speed is on the priority list for the next release.

I don't understand why Select by Location would use so little CPU and why is it performing I/O at all.

It is using a file based database, hence the occasional disk access.