When I first put my library in Calibre I had something like 3000 duplicates showing. As the library was over 30000 records, that was not unexpected.
To narrow it down started with first looking for identical (binary) which was an automatic delete one of them.
Then looked for identical author, identical title and chose the best formatted OR if it was two differing formats which I wanted to keep both of added a note to explain why (for instance two different translations of the same book).
That should reduce the duplicate records to a manageable number which you can go through at your leisure. It took me about two months of an hour here, an hour there to clean up my duplicates completely.
|