[Koha] Koha Digest, Vol 56, Issue 13

[Stacy Pober]

Hi Jos,

We're using a Koha-like system (Liblime Enterprise Koha, a/k/a LEK),
but I think our method would work for you.  I've added a lot of ebooks
to our catalog.

Before I upload the records into the catalog, I pre-process them in
MarcEdit.  MarcEdit is a free program you can download from:
<http://people.oregonstate.edu/~reeset/marcedit/html/downloads.html>
or <http://bit.ly/9SOCay>

With MarcEdit you can do bulk operations to add fields, delete them,
and edit them along with a number of much more sophisticated
functions.

Install MarcEdit, and open your MRC file.  To add a field, you go to
Tools->Add/Delete Field

In the first box, you type the field name, in the second, you put both
the MARC indicators and the subfields.  For our ebooks, the 952 data
looks like this:
\\$aMAN$bMAN$cEBOOKS$oELECTRONIC BOOKS$yEBOOKS$8EBOOKS

You can find information on these subfields in the Koha manual at:
<http://koha.org/documentation/manual/3.0/cataloging/item-fields-data-migration>

Although the books are electronic so there is no physical copy, we
still fill the subfields for location, sublocation, and call number
because we ran into problems when we omitted the $a and $b fields (and
you can read about the questions I asked about those fields  in this
list's archives, I think.)

Here are some things to watch out for when processing ebook records :

- check that the ebook vendor has given you clean records.  I noticed
that some of the vendors are not vetting the records properly, and are
including 856  fields (Electronic  Location and Access) with URL's for
other ebook vendors in their records.  So there will be the good link
to their subscription service and another link to another ebook
vendor's version to which that your patrons will have no access.

- use the MarcEdit Field Report tool to see there are no anomalies in
the records that you might want to clean up. Some of the vendors sent
records with occasional 5xx and 9xx fields with data specific to some
other library.

- consider adding a field with a unique identifier for that ebook set.
 We add a MARC 500 field with the name of the vendor or the particular
book set so that if we want to search in the opac for just a book from
that set, we have a way to do that.

- consider if there are any fields you want to delete from the record.
One of the vendors added four separate fields to each record which
happen to mention their company's name.  Some of those fields were
appropriate, but some seemed to duplicate other fields, so we just
removed them.

- if you are using the XLST display, then the icons will be mapped
primarily to characters in the MARC leader, and will NOT be determined
by your ITYPE or CCODE designations.   This may cause some confusion.
We found that some vendors correctly changed the leaders to show that
these were books in electronic format.  Springer Verlag uses leader
codes that indicate that the books are actually non-text computer
files. Another ebook vendor used leader codes that erroneously
indicated that the ebooks were actually print books in paper format.
You can change the leader codes in MarcEdit if there's a problem and
you're using the XSLT display

- use the MARC Validator tool in MarcEdit to look for errors in the
records. One warning about that, though:  the tool is extremely
strict, so you might want to go into the marcrules.txt file that it
uses and actually loosen the validation rules.  A specific example is
that many MARC records from ebook vendors include a notation in the
020 field as to whether the ISBN listed is for the paperback,
hardbound, or electronic version.  These will generate an error using
the default marcrules file for validation, so I changed that rule.

- if you use a proxy server such as EZproxy for off-site access to the
ebooks, you can use MarcEdit to add the proxy server prefix to the URL
field $856$u of each item.  You would use Tools->Edit Subfield Data
for that.

MarcEdit is a wonderful tool for batch editing MARC records.  It's
powerful and flexible and there is a nice email listserv for support.
It's not open source, but it is free.  There are youtube videos that
Terry Reese has put up to help people learn how to use the program.
It's a reasonably user-friendly program, but for mastering some of the
functions, I found the videos very helpful.

After you've made the changes you want to the file in MarcEdit, you
can upload the batch with the new 952 fields, etc.  and they should
not have the same problems you're seeing now.

You may find that your ebook vendor gives you updates that are
additional records, but some of the vendors also do deletions, which
gives you two choices.  When your collection changes, you can either
use the Undo Import function to remove the initial upload, then upload
the new version.  Or, you can add the new titles, and then manually
delete any records for books or editions that  the vendor is no longer
offering.  I think using the Undo Import and just re-loading the new
batches is the easy course for updates, but unfortunately, we will not
have that as an option.

LEK is in constant development and testing mode and here is some new
software in the next release that did not work well in testing with
the batch upload "undo" feature that is standard in open-source Koha.
The vendor is removing the "Undo Import" function as a result.

Since most of the books I uploaded were on subscription rather than
permanent purchases, this means I'm going to have a lot more hassle
when we need to remove the batches.  Without 'undo import' removal
will require a special request to LEK support for manual removal, and
I can't even contact them directly. So, I'm going to miss the undo
import feature. Ebrary supplies the largest number of ebooks to us and
they recently contacted me to say that they'd fixed some problems I
reported and now have improved versions of their MARC records
available for download (there were bad diacritics in the batches we
used).  In the past, I would have taken out the old batches and
uploaded the new versions, since it was so easy.  Now, I'll probably
stick with the version we have since removing batches is not available
to LEK end-users.

Since you're using the free, open source version of Koha  so if you
need to update the ebook holdings you can just undo the old batches
and replace the records pretty easily.

One more thing: if you are replacing an old batch of ebook records
with an updated file, I would not use the "record matching rule" to
replace old versions with new ones.  Unfortunately, some of the ebook
vendors do not supply unique ISBN numbers for their electronic version
of a paper book. If you use the ISBN as your matching rule and you
have a record for the paper version of the book already in the
catalog, you may overwrite that record with the ebook record.

I hope you find this information helpful.

-- 
Stacy Pober
Information Alchemist
Riverdale, NY 10471
stacy.pober at manhattan.edu

On Mon, Jun 7, 2010 at 4:10 PM,  <koha-request at lists.katipo.co.nz> wrote:
> Message: 3
> Date: Mon, 7 Jun 2010 11:08:57 +0100
> From: Jos Black <jos.kohado at gmail.com>
> Subject: [Koha] Items for bulk ebook MARC records
> To: koha at lists.katipo.co.nz
> Message-ID:
>        <AANLkTikzbbvQkQ3izoKETFplfmuzjTO53OXxd5XesU2F at mail.gmail.com>
> Content-Type: text/plain; charset="windows-1252"
>
> Hello
>
> This is my first question to the list since introducing myself a few months
> ago,
> so please forgive me if I?m asking a silly question, but I haven?t yet been
> able to find an
> answer out there.
>
> I?m currently trying to set Koha up as a catalogue and access gateway to our
> eresources collection, as our current (proprietary) library system is a
> nightmare to import anything into.  As a first attempt I have successfully
> imported nearly 3000 MARC21 bib records from one of our ebook suppliers into
> a
> test install of Koha (3.00.00.107, hope to upgrade soon!).  However, when I
> search for any of the titles in the OPAC, I get the ?Availability? message
> ?No
> items available? in the brief results, and in the detail holdings tab I get
> ?No
> physical items for this record?.   The MARC records as supplied don?t have
> any
> item data, either in the Koha 952 item tag, or even item data I could map
> across
> in the 852, and I?m yet to hear from the supplier that they can supply item
> data.
>
> These messages will be off-putting for most of our readers, many of whom are
> new
> to using libraries, and the vast majority probably won?t ask for help, so I
> need
> to find a way to sort this out.
>
> How could I tackle this apart from a huge effort in reader education,
> modifying
> my copy of Koha, or individually creating thousands of items?  I have a
> couple
> of ideas but don?t want to reinvent the wheel so would truly appreciate your
> help.
>
> Thanks!
> Jos