Length limit in Marc21

I've become older and wiser investigating something that puzzled me.

I'm looking at this record
librarycat.org/lib/bnielsen/item/101634935

The MARC view for that shows a very long field 921 a

I want to convert the MARC record to Marc21 (to be able to use another tool called marcli) but Marc21 is format where the length is given by four decimal digits (i.e. maximum is 9999 characters).
My current workaround is to chop of the field if it is too long.

I wonder if anyone here could comment on other solutions? I.e. splitting into several 921 a fields?

Is there a practice for handling that limit in Marc21?

As far as a system like Voyager and Alma, yes. A cataloger would break up the 921.

Looking at the field, that's terrible cataloging. It's just a string of sentences with no context.

>2 lesmel: It's the review field from the LT record so it's just taking that and throwing it into 921 a
So it's me, that's a terrible cataloger :-)

At job I deal with people who confuse Title with CV and use Building/RoomNumber to communicate that they are on leave so it's just a kind of revenge.

Thanks for the input! I'll chop up my lengthy reviews into less lengthy clumps of 921 a's. (Indeed chopping at the first space character after 9000 characters seems to work nicely.)

And just to help others with the same problem, I'll note here that the error messages produced by yaz-markdump are a bit hard to decipher when you give it a record with a field larger than 9999 characters.
So maybe that's a useful hint to anyone searching for a mention of yaz-markdump here on LT.

A bit more on what I'm doing. The idea is to take the TinyCat MARC view of my books. That's in html, but a script can convert that into a text version of Marc and break very long lines. I can then correct the many character set issues in the records and convert them into Marc21 format. And then use a tool called marcli (i.e. marc listing).

Here's a link to the Marc21 format
https://www.loc.gov/marc/bibliographic/bddirectory.html

An example of using marcli:

./marcli_linux -file lt.marc -match 'Broadhinny' -fields 001,245,100
=001 39269895
=245 1\$$aDommen er afsagt : $bMrs. McGinty's død / $c(Overs. ved Vera de Journe)).
=100 1\$aChristie, Agatha.

Field 001 is Book_Id making it easy to find the book in LT.

And something more useful (since I didn't know that the books were in a series: Svanebøger.

./marcli_linux -file lt.marc -match 'Jukan' -fields 001,830
=001 205709597
=830 0\$$aJukan ; $v1.
=830 0\$$aSvanebøger.

=001 206120864

=001 206138391
=830 0\$$aJukan ; $v3.
=830 0\$$aSvanebøger.

=001 206138692
=830 0\$$aJukan ; $v5.
=830 0\$$aSvanebøger.

=001 206139336
=830 0\$$aJukan-Serien ; $v6.
=830 0\$$aSvanebøger.

=001 206139634
=830 0\$$aSvanebøger.

=001 206140285

=001 206140292
=830 0\$$aJukan-Serien ; $v12.
=830 0\$$aSvanebøger ; $v10.

=001 206140506
=830 0\$$aJukan-Serien ; $v8.
=830 0\$$aSvanebøger.

=001 206140529
=830 0\$$aJukan ; $v4.
=830 0\$$aSvanebøger.

...