Jim,
Are you saving this data back to Connexion? If so, leaving the 008
should be fine.
If you're saving the file external to Connexion, then the 008 needs
changing.
I've been able to make the "=008 20" with "=008 " replacement with no
problems in the past, but my conscience said OCLC should get the export
credit, so now I highlight all records in local save file and do a batch
export to a file on your PC. This will cost your institution the OCLC
export charge, but will convert the 008 to the correct format.
If you've already done significant editing to the original file, you
might be able to merge the new file's 008 into the first one using
Terry's wonderful Merge tool.
My $0.02
Tim
Timothy M. McCarthy
University at Buffalo
716.645.8577
[log in to unmask]
On 2/8/2011 6:52 PM, Jim Kuhn wrote:
> Hello:
>
>
>
> I recently pulled in some OCLC local save file records and it took some
> weeks before we noticed that the Connexion bib file reader was
> interpreting the year portion of Date Entered (positions 00-05) as yyyy
> rather than yy. This pushed our 008 data off by two bytes. For example,
> compare the first 008 below (which is fine) with the second (which
> isn't):
>
>
>
> 090924 m 1577 1583 xx_ ad__ _ _ ____ _ 0 0 0 _ 0 _ eng
> _ d
>
> 200909 2 4m15 7715 83x x_ad _ _ ____ _ _ _ 0 _ 0 _ 0_e
> n g
>
>
>
> The thing is, the apparently-missing final two characters are actually
> present. So I've been puzzling over how to delete just the 1st two
> characters of each 008 in affected records. That is, I want to make 008s
> that look like this in MARCEdit:
>
>
>
> =008 20051110s1603\\\\xx\n\n\\\\\\\\\\\nknger\d
>
>
>
> instead look like this:
>
>
>
> =008 051110s1603\\\\xx\n\n\\\\\\\\\\\nknger\d
>
>
>
> Here's what I've tried, despite knowing only enough about regular
> expressions to be mildly dangerous:
>
>
>
> - MARCValidator reports these 008s out as too-long. Right.
>
> - A straight replace to change "=008 20" to "=008 " corrupts
> the file. Although I successfully produced an uncorrupted marc file by
> manually modifying a single record, there are thousands of records
> involved. Haven't yet tried my fallback method yet: write a macro to
> just hop from 008 to 008 "manually" editing each in turn. Inelegant. And
> maybe not effective? But I'll try it if I can't do this globally.
>
> - Edit subfield data, to find "([0-9]{2})(.*)" and replace with
> "$2". Tried this using both "remove" and "replace", and by calling the
> 008 position "00" or "02" or "00-05" or "05". No luck.
>
> - Then I thought I'd try regex on Find / Replace, or the Script
> Wizard. But I don't do a heck of a lot with regex, and before I start in
> on it I thought I'd check in here. Besides its not clear to me that 008s
> are even editable in this way?
>
>
>
> Thanks in advance for any advice re regex in Replace, in Edit Subfield
> Data, or in the Script Wizard. Or any other method for deleting just the
> 1st two characters in invalidly-long 008s.
>
>
>
> Terry if you'd like a closer look at how the Connexion plugin is
> behaving with this data I can send you our source bib.db.
>
>
>
> best,
>
> Jim Kuhn
>
>
>
> __________________________________
>
> Head of Collection Information Services
>
> Folger Shakespeare Library
>
> 201 E Capitol St SE
>
> Washington DC 20003-1004
>
> 202-675-0334
>
> [log in to unmask]
>
> www.folger.edu
>
>
>
>
> ________________________________________________________________________
>
> This message comes to you via MARCEDIT-L, a Listserv(R) list for technical and instructional support in MarcEdit. If you wish to communicate directly with the list owners, write to [log in to unmask] To unsubscribe, send a message "SIGNOFF MARCEDIT-L" to [log in to unmask]
>
________________________________________________________________________
This message comes to you via MARCEDIT-L, a Listserv(R) list for technical and instructional support in MarcEdit. If you wish to communicate directly with the list owners, write to [log in to unmask] To unsubscribe, send a message "SIGNOFF MARCEDIT-L" to [log in to unmask]
|