On Thu, Feb 24, 2011 at 4:09 AM, M. Brooke Helman <abesottedphoenix@yahoo.com> wrote:
And I would like to "add value" to them later on - say I find a MARC
Salve! source that assigns subject headings in a way that I like, or adds links to cover images, etc. looking to selectively overlay/insert a subset of the new records into my existing ones.
How does Koha handle this?
Okay, that's a good question :)
Koha pretty much doesn't right now, but folks are taking some hammers to the code to make it try and manage that safely. PTFS kind of doesn't get that you need a scalpel and not a caber for this job. (What else is new?)
Or would I be better off exporting my current records to MARC, using an external MARC editing tool to handle the merge and then re-import? Can anyone suggest tools for me to look at? I'm aware of biblios.net and MarcEdit so far.
Currently yes, and I'd use the latter not the former, since the former sucks.
Or if it's really better to do this one record at a time, does that change the recommendation as to using Koha for the process?
Mebbe. Sometimes Koha can put weird stuff into a MARC record, though that's mostly over with by now. I used it as an editor since I was in the middle of nowhere and no one would ever notice the aforementioned weird stuff anyhow.
And finally, what if there are certain field values that I do want overwritten in addition to the new ones being brought in, but some fields shouldn't be touched at all?
Hmph. I'd wait for Nicole's word on that.
Cheers, Brooke
Thanks Brooke. I'm thinking I'll use Koha for one-at-a-time editing, and try to develop a sound workflow for bulk updating using external tools. Here are my current ideas - sanity check (from anyone) appreciated - and I realize it's a kludge, wish I were up to speed on text tools like Perl 8-( Unfortunately it seems that MARCedit is good for automating merges when you're merging in specific fields from consistent sources, matching control numbers etc, not for the kind of ad-hoc updating I'm doing. Note the below assumes I'm collecting a bunch of MARC records from a variety of sources, the goal being to cherry-pick the best quality data from each - some sources have good Subject headings, others more accurate imprint/collation data etc. 1 Export MARC records from Koha with a solid local control number (001) 2 Split these out into one-biblio-record per file named "koha.mrc", each in a folder named with the control number, drop item-level fields. 3 Bring in external MARC records, each file downloaded to the per-biblio folders, files named after the source. 4 Use MARCedit's to join all files per folder into one working multiple-record MARC file per title, then export as CSV and bring into a spreadsheet. One worksheet per title, one record per source MARC file. Column headers show the MARC fields used, everything's all in one place easy to scan. 5 Copy and paste, consolidate all into a single row (ensuring koha control number remains), delete the fields not used in Koha. Save/export the new master record back as MARC into the same folder named "new.mrc", 6 Join all the new.mrc's into a single file, stage into reservoir. 7 Use (one of the) Koha merge functions to update the existing biblios based on the control number matching - all existing data in each record overwritten by the incoming data. Obviously it would be great if either Koha or some other external tool allowed one to display multiple MARC records on screen in a GUI, allowing one to tick off the fields containing desired data, copying and pasting as needed and then one-button "merge" as specified to produce a new record. But it's a lot to ask I know, maybe one day Koha will have some of this. If I can help the developers at any stage from a motivated user's POV, someone please let me know how. And thanks in advance for any feedback on the above, especially suggestions on how to simplify a rather convoluted process, this just doesn't seem to be an area where one can KISS. I wish I could just say "eh good enough" and move on, but my OCD perfectionism won't let go 8-)