New Ingram Import

[rivka, Former Moderator]

Earlier today, we started importing a new batch of Ingram imports. We have modified our filtering based on feedback from previous imports, but no filter will catch everything.

Please let us know if you see any unusual data associated with the data source "Ingram" with dates on May 14, 2012 or later.

[Lobstergirl]

Can you explain what "modified our filtering" means, and give some examples?

[Lobstergirl]

I wonder if it's possible to have an Ingram import ignore ISBNs already in the database, and only import non-existing ISBNs. That would certainly cut down on the Ingram errors.

[Chase]

One major change is that we aren't adding new contributors to existing books, since that seemed to cause a lot of trouble in the past.

The importer also attempts to remove author names that appear in book titles spuriously. For example, before this import began, the title of this book was "Cake Angels: Amazing Gluten, Wheat and Dairy Free Cakes. by Julia Thomas" [sic]:

http://www.goodreads.com/book/show/12...

However, as rivka noted, there are many different formats in which this error appears, and we can't catch all of them automatically.

As for limiting the import to new ISBNs -- we use book data from a variety of sources, some of which are more reliable than others, but all of which we need in order to offer the most complete book data possible on Goodreads. Imports must be able to update existing books in order to verify or correct information supplied by other sources.

Sometimes it may seem like imports do nothing but introduce mistakes; however, it's important to remember that the squeaky wheel gets the grease. It's easy to notice a thousand mistakes while losing sight of the hundreds of thousands of correct updates that a single data import can provide. Librarians are vital to Goodreads, and we want to use their time wisely, so we let the import do all the heavy lifting, knowing that the cleanup work is relatively little. Not little, we know, but relatively little.

As always, once librarians update a book's data, that update takes precedence over (and will not be overwritten by) any future imports that may occur.

[Lobstergirl]

Thanks.

[The Elusive]

rivka wrote: "We have modified our filtering based on feedback from previous imports"
That made me smile. So formal-sounding. Did you fix the "Spanish Title = English title" issue? Being Spanish I find that one a lot.

[Sarah (Presto agitato)]

I'm not sure if this is the right thread to post this, but I'm not sure why Ingram is changing the capitalization and spacing of the title of this book incorrectly. I changed it back, but I'm wondering how many of these are out there. I only just noticed this one.

http://www.goodreads.com/book/edits/2...

[rivka, Former Moderator]

Individual books like that need to be fixed by librarians, but there's not much more that can be done. We're looking for patterns, and one book does not a pattern make.

For every book the Ingram import made the capitalization worse on, there is probably at least 2 or 3 that it made it better.

[Ellie]

In this case, a librarian page watch option would be very helpful.

[Lobstergirl]

Perhaps "prepak" and "prepack" should have been filtered out. These words would indicate something needing to be nabbed about 99.999% of the time, whereas another word like "display" would cast too wide a net.

A search for "prepak" for example finds 759 items, every one I've looked at imported the first week of June.

[rivka, Former Moderator]

The filter we implement to catch those was added June 5.

[Lobstergirl]

Ah, okay. Makes sense now....

[Lobstergirl]

Actually, I take that back. I'm seeing prepaks from June 6-7.

[rivka, Former Moderator]

If we need to refine the filter, it helps to have examples. Links, please?

[Lobstergirl]

From 6/6
http://www.goodreads.com/book/show/15...
http://www.goodreads.com/book/show/15...
http://www.goodreads.com/book/show/15...
http://www.goodreads.com/book/show/15...
http://www.goodreads.com/book/show/15...

6/7
http://www.goodreads.com/book/show/15...

6/8
http://www.goodreads.com/book/show/15...
http://www.goodreads.com/book/show/15...
http://www.goodreads.com/book/show/15...

6/6 (clip strip)
http://www.goodreads.com/book/show/15...

6/8 (clip strip)
http://www.goodreads.com/book/show/15...

[rivka, Former Moderator]

Great. I've added those to the ticket.

[Linda]

Have failed miserably at every attempt to use this website. Trying to get my book Unintended Lies by Linda Kendall McLendon posted. Rejects my e-mail name, rejects my password, I have changed passwords 5 times in one trial. Everyone raves about this site and I'm perplexed and a new author to boot! Help!

[rivka, Former Moderator]

Have you tried posting a new thread in this folder: http://www.goodreads.com/topic/group_... ?

[Lobstergirl]

"27cpy" should be filtered out. A search produces 167 results.

http://www.goodreads.com/search?utf8=...

[rivka, Former Moderator]

Added that one to the ticket as well.

[Lobstergirl]

Also "24 copy" - 182 results.
http://www.goodreads.com/search?utf8=...

"24cpy" - 112 results.
http://www.goodreads.com/search?utf8=...

"30 copy" - 28 results.
http://www.goodreads.com/search?utf8=...

"36 copy" - 78 results.
http://www.goodreads.com/search?utf8=...