Sorry to hear about the problem – could you confirm which version of Mendeley you are using? (You can check via ‘Mendeley Desktop’ -> ‘About Mendeley Desktop’ in the menu). Can I also check whether Mendeley tries to re-import files on every startup or only in certain circumstances (eg. clicking the Recover button in the Crash Recovery dialog when restarting Mendeley after a crash)?

Mendeley Desktop should be checking for duplicates in several ways during import but evidently something has gone wrong. I should clarify what is supposed to happen here:

1. When Mendeley imports a file via a watched file it records the filename in the MonitoredFiles table of a sqlite database called monitor.sqlite in the Mendeley data directory. When Mendeley starts up and scans your watched folders, it should skip any files whose paths are already in that database.

This file does currently get deleted though if you click ‘Recover’ if prompted by a crash recovery dialog – the idea being that if the program crashed when importing a particular PDF or batch of PDFs it should avoid trying to re-do the import straight away on the next startup.

2. When you import a file via any means (eg. drag-and-drop, File -> Add Files) Mendeley will check if that file is already associated with any of the existing documents in your library – in which case you’ll still see the ‘Importing X files’ message in the status bar but it should just skip that document once it realises that the files are the same.

3. Finally there is a slightly fuzzy metadata comparison which compares the imported metadata against those of articles already in your library for duplicates.

Thanks for the feedback – you can get in touch via support@mendeley.com (or my address here) if you have any questions.