I believe IMDb has 2 different formats they export and we support both of them. So unless there's a third format that I'm not aware of you should be good to go.

Thanks, Travis. I'll give it a try then.

It looks like it worked! Well, more or less... Some 600 ratings got lost in the process. It makes sense for about 50 of them (TV series episodes, unknown shorts, ...), but I can't figure out what happened to the others, nor which titles are missing.

If you head back to the import list section, you can view your import history, and on the list, it will show which items didn't make it.

Yes, I can see why the system kept 5,131 movies(+ratings) out of the 5,145 it detected. But I still don't understand why it detected only 5,145 out of the 5,847 in my file (which is more like 700 missing, actually ).

Ah, I understand now. It's that they weren't detected in your list in the first place.

When I wrote the importer and tested a bunch of different files I learned that IMDb's files are not very well formatted and have a lot of little issues that make it very hard to build a parser that handles everything. There were some cases that I came cross where I simply had to skip the row because the CSV formatting was so wonky. I would suspect that's the problem.

Yes, it's probably the character encoding standard used for IMDb's export file that's to blame. I've checked a few movies with non-standard characters in their original titles: none have been imported.

Sorry for my meddling.
Maybe a RegEx? Leave the Title and the name of the Director blank?
What are the required fields for an import?

position,const,created,modified,description,Title,Title type,Directors,You rated,IMDb Rating,Runtime (mins),Year,Genres,Num. Votes,Release Date (month/day/year),URL
1,tt1060277,Thu Apr 9 00:00:00 2009,,,Cloverfield,Feature Film,Matt Reeves,8,7.1,85,2008,"action, sci_fi, thriller",240707,2008-01-16,http://www.imdb.com/title/tt1060277/
2,tt0409459,Fri Mar 6 00:00:00 2009,,,Watchmen,Feature Film,Zack Snyder,9,7.6,162,2009,"action, mystery, sci_fi",315606,2009-02-23,http://www.imdb.com/title/tt0409459/

I only use the IMDB id, rating and date. The problem isn't the actual data I try to import, it's the CSV parsers ability to iterate over rows because of badly placed and quotes and escape characters and the like. In my case, I am using Ruby's standard library CSV parser.

I managed to make it work! Here's what I had to do:

1. export ratings from IMDb ratings.csv file
2. edit that CSV file with Excel, to keep only Const,Your Rating,Date Added columns excel.csv file
3. replace all ";" with "," using the notepad (it seems Excel turned the "," into ";" when I saved the excel.csv file) final.csv file
4. import final.csv file into TMDb

Good job Francis! It doesn't surprise me that a solid commercial product like Excel could clean those files up. That's a very useful tip to know about. Thanks for looking into it some more.