This topic is currently marked as "dormant"—the last message is more than 90 days old. You can revive it by posting a reply.
Warning: simplexml_load_file(http://sru.kb.nl/sru/sru.pl?x-collection=GGC&version=1.1&operation=searc...): failed to open stream: Connection refused in /var/www/html/inc_externalbooksearch.php on line 115 Warning: simplexml_load_file(): I/O warning : failed to load external entity "http://sru.kb.nl/sru/sru.pl?x-collection=GGC&version=1.1&operation=searchRetrieve&query=isbn+all+9027418721&recordSchema=dc&startRecord=1" in /var/www/html/inc_externalbooksearch.php on line 115 Fatal error: Call to a member function children() on a non-object in /var/www/html/inc_externalbooksearch.php on line 119
If you want to reproduce the error, the ISBN I'm searching is 9027418721.
The problem is that LT queries SRU. The KB doesn't use SRU anymore, but now uses JSRU. The query should be formatted as such:
When I run the query above from Aryanhwy in that manner I still don't get the response I expect, but mine works. It appears that the part of query (isbn+all+) is too much. So this works:
version=1.1 also doesn't seem necessary for the query to work.
So this should help the LT wizards along, but like I said, I am by no means an expert or officially saying all this.
If you add: &x-fields=ISBN to the query, a field is added to the XML output to identify the ISBN. Otherwise the ISBN is one of the dc:identifier fields, but you can't see which one.
And a request of my own, it used to be (with SRU) that when you added a book from the KB database, it never took the year (year published was always ?). Is it possible to fix this too, using the dc:date field?
1) you can search for Titles, Authors etc. by putting the query between "". In that case ISBN-search still works.
2) for books originally published under ISBN-10, it's only possible to search by ISBN-10. But with a small tweak it is also possible to use the barcode: strip the first 3 digits of the barcode, and replace the last digit by an *, so barcode 9789018017255 becomes query 901801725* (the ISBN is 9018017256).
(books originally published under ISBN-13 can only be searched by ISBN-13, which happens to be the same as te barcode.)
3) the best dc:title-field is the one with the most information, but you have to strip everything behind the / (which are only credits).
So the following URL seems to be working for most books:
Soon I hope to post or send you a more detailed overview of the returned xml.
UPDATE: I sent the xml-analysis by email to librarythings info-address
Even though this bug is apparently is solved, it is not for me. Even with the above tips, I draw a blank every time.
Being lazy, I like to just be able to type the name of the writer and title of the book, and get several hits. At the moment, all I ever get when trying to add books is : Niets gevonden.
I still have hope that they (LT) fix the import of language and year import from the data received from the KB.
I would like to see more books, when I only gave the name of the writer in the search box from LT.
When you ask for this kind of informations at the Vlaamse Bibliotheek I get serveral pages, with lots of titles. The information from the KB is only 1 page maximum.
None of the experts are in this week, so I don't know the exact query, but I will ask next week when they are back in.
I will also ask about the site map/catalog difference.
miketopper, can you please look at this again? Please?
Our thinking goes something like this:
1. The non-MARC sources can't be maintained. They represent entirely separate processes for getting and munging the data. Once we get it, the data is too different, so we can't really revisit it as needed, the way we've done when we've added physical description and etc.
2. We can add a few other Dutch libraries we've never connected to before.
Losing KB would be a big loss, I understand, but I think it may be necessary for moving forward with Add Books generally.
I just tested the ISBN I had entered 9783258069937 and it finds the book in their database.
Ok, I asked the guys at work again, and this is what the expert says (translation to English by me):
- KB uses CQL as the query language. It is not supported completely, but the basics are supported.
- The ZING library is used to parse the queries
(I have no idea if this is imported to LT support, but I included it anyway)
If you want a query that searches multiple fields, like in the example Harry Mulisch Vrouwen (and expecting to find Twee vrouwen, this is the query to use:
Even more specific is the query:
But that searches in specific fields and I don't think that is how Add Books works on LT. It does get less hits because the dc:description field is not searched.
I also mentioned the SRU/MARC debate/issue/whatever that has been mentioned here and in the other Dutch Sources topic. That is of course a bigger thing, but it has been noted.