The series search shows all series pages with the search term

From http://www.librarything.com/topic/102604#2332965

So, what's left, anyone?

Something which is likely to become increasingly annoying as time goes by:

The series search shows all series pages with the search term, including ones which have no works. For example, I used the series page to find and consolidate duplicate 'Eyewitness' series http://www.librarything.com/search.php?term=eyewitness . Three of these series pages are now empty:
http://www.librarything.com/series/Eyewitness+travel+guides
http://www.librarything.com/series/Eyewitness++Travel+Guides+
http://www.librarything.com/series/DK+Eyewitness+Top+10
but they continue to show up in the search. Over the long term, this will lead to a lot of unnecessary checking, and wasted effort.

What this is actually turning up is CK data that's "hidden" in a work that's been combined with another work. I actually think it'd be useful to have this turn up, *if* there's a way to remove the data. But as long as we can search for and find the hidden info *somewhere* (and please don't get rid of the existing search mechanism in CK history, since it does slightly different things), I guess it's ok to get rid of it in the main search results.

For example, for Eyewitness travel guides, there is one book with that listing:
http://www.librarything.com/commonknowledge/changelog.php?item=3608077&type=...

When it was combined with another book, the CK stayed with the original work number, but no longer displays anywhere except in CK history.

#2
What this is actually turning up is CK data that's "hidden" in a work that's been combined with another work

I don't think it just that, though. At least one of those series was one I created myself, by mistake (a matter of copy-and-paste where I'd included an extra space without noticing.) I'm pretty sure that in that case, I edited all the incorrect data.

My understanding (and the reason why I brought it up) is that once a series page is created, the page header stays even if all the works are removed again. It appeared to me that the search was finding and returning all these pages, all well as ones populated with works.

4> once a series page is created, the page header stays even if all the works are removed again

hasn't this been the case for a while? and with most CK? I seem to recall seeing a lot of not-actually-there CK show up in the drop-down suggestions for quite some time.

4-5 I'm not sure that's the case, though. I've managed to get rid of some "blank" awards pages from the search results by going in and deleting the "hidden" layer data. (There is a way to do it which is a very complicated manipulation of URLs, plugging in all kinds of variables and numbers, but it's a big dangerous to use because it can delete everything on the wrong work if there's a typo.)

Off topic a little bit, but...

rsterling: "and please don't get rid of the existing search mechanism in CK history, since it does slightly different things"

Can you talk about that a little more? We're talking about rolling that into the site-wide search mechanism very soon and it would be good to have some members' views on how they use the current CK search.

I use the existing CK search to find every *instance* of a term, rather than the term itself. I want something that turns up hundreds of records, date-stamped, for whatever term I'm searching for, NOT just a list of series pages, or awards, but a list of every book or author that has that term in the award or series field. I also use the CK search to search for and find changes made by a particular user. I like the fact that it's a search of the history, and that the search results display in the history, with all the history columns.

To give an example relevant to this thread: I want a way to find "hidden" data. So if I search for Spam in the series field, I want a way to find all the records that have that word in the series field, even if they've been combined into other records and the Spam series page is thus now empty. I want a thorough search that shows me *all* the entries, not just the ones on the front-level of data.

Another case where this is useful: when 2 works get combined, their CK gets combined. However, sometimes things get combined incorrectly, and so it's useful to search CK to figure out what was attached to the original record, pre-combination, and thus what needs to be moved once the works are separated out. If a Tolstoy book has wrongly been combined with a Dostoevsky book, and suddenly has CK data for the Dostoevsky book, I can search for a piece of that data, find both its instance in the new combined work and its instance at the old work number for the Dostoevsky book, and then look at the history for that old work number, to see what the CK was that needs to get moved out of Tolstoy.

Update on why this isn't fixed yet (see http://www.librarything.com/topic/107331).

Need to look into this. Sounds like a 1-2-hour rabbit hole.

Dupe bug report (by me):

http://www.librarything.com/topic/114203

Comments Tim left on my dupe report make me think this is fixed now. I tried the example in #1 and those old series weren't showing up.

I'll leave it to Tim to actually mark as fixed. Just wanted to bump it up to the top to get his attention.

Thanks. Fixed.