And research this morning has me on lots of other libraries' websites running null searches against their "articles" database and seeing what their total results are. 1,000,746,600 at Duke. 1,000,324,214 here at PSU. Princeton is squeaking UNDER a billion with 998,667,021... this is utterly unsustainable, tbh. This is helping shape my decisions. #library #libraries #research
I'm going to be making a decision which will dramatically filter initial search results to resources which are either scholarly or peer-reviewed and NOT book reviews (because we have 9.8 million of those and the system we subscribe to does NOT do a good job putting the real book above them) ... this brings us down to 88 million records. A lot, but that's a TON of article records. 83.6 million, to be almost-exact. So I think that's ok.
@platypus We're enough smaller than y'all to not get quite such large results, but still it's ridiculous. Just makes people's eyes glaze over, at best.
Good on you for making that change. I've not gotten to a place where I'm comfortable enough with Primo to attempt something like it.
@platypus *nods* That's how we have it set up now, but wrt availability. There's a little checkbox to expand the search to "outside the library's collection".
Out of sheer curiosity, what search term/s got you that many hits in the first place?
@auntiekiki null search. just hit enter and go! But now... and results beyond the library's collection is 1,693,019,077 so that's def not the number either. I just emailed our ER person to ask about it because it's possible we had a mid-month package shift, though 378 million records is a lot to... drop? I'm getting the comparatively-sane 621,959,034
@platypus Huh. I can't do that. Have to put in some bit of text or get a "no results found."
@auntiekiki yeah, this is a lucky? thing with Summon. Princeton's: 998,667,021 results. Duke: 1,000,746,600. Yale: 817,076,655. Dartmouth: 878,541,266. Columbia: 1,051,600,796
...Stanford won't let me do it because they Blacklight their Summon goshdarnit. Brown/Smith/UIUC use EBSCOhost.
Or you can do something like search "skeleton" and then remove the word (in some infrastructures) e.g. click the X here: http://search.library.yale.edu/articles?s.cmd=&s.ff%5B%5D=ContentType%2Cand%2C1%2C20&s.ff%5B%5D=Discipline%2Cand%2C1%2C20&s.ff%5B%5D=Language%2Cand%2C1%2C20&s.fq=&s.ho=t&s.q=skeleton&s.role=authenticated&spellcheck=true
@platypus Hrm. Guessing ExL doesn't want a null search possible due to server load or something (we're hosted). No way to remove the word afterwards, either.
So far, largest set I've been able to return is for "red", at 14.5 million (searching outside lib's collection).
@platypus Oh, I forgot the obvious: the. Just shy of 222mil. (Surprisingly, "a" was less, at ~155mil.)
@auntiekiki huh! I wonder how many other things would be turned up, though, because that still seems a wee bit low. Our ER person says "yep your first number sounds right, the second sounds wrong" so... we're gonna keep an eye on that.
@auntiekiki one critical aspect is people will still be able to do things like unfacet and decide to search for just Book Reviews if that's what they're actually after.
...related, I'm NOW having trouble replicating the 1,000,746,600 result which is the most Summon thing tbh