Changes to sorting by 'Call Number' in Zotero 6.0.5
One of the entries in the ChangeLog to Zotero 6.0.5 is 'Improved call number sorting in items list'.
This is certainly the case with LoC call numbers in respect of the so-called Cutter fields which are now collated accurately, i.e. PJ4519 .B9798 A58 1999 now correctly precedes PJ4519 .B99 A65 1976; i.e. Cutter fields are collated decimally such that the .B99 in the second example is effectively right-zero-padded to .B9900 before sorting.
However, the ordering of classification fields per se, i.e. from the beginning of the call number up to the next '.' followed by a letter, now also seem to be treated in this same way rather than the previous (and, as I understand it, standard) alpha-numeric ordering. The result is that PJ403.B64 C666 now wrongly follows PJ3930.S49 A53 2015; i.e. all PJ3*s before PJ4*s and all PJ4*s before PJ5*s, and so on. Should there not be a continuous numerical sequence under each classification?
Thanks
This is certainly the case with LoC call numbers in respect of the so-called Cutter fields which are now collated accurately, i.e. PJ4519 .B9798 A58 1999 now correctly precedes PJ4519 .B99 A65 1976; i.e. Cutter fields are collated decimally such that the .B99 in the second example is effectively right-zero-padded to .B9900 before sorting.
However, the ordering of classification fields per se, i.e. from the beginning of the call number up to the next '.' followed by a letter, now also seem to be treated in this same way rather than the previous (and, as I understand it, standard) alpha-numeric ordering. The result is that PJ403.B64 C666 now wrongly follows PJ3930.S49 A53 2015; i.e. all PJ3*s before PJ4*s and all PJ4*s before PJ5*s, and so on. Should there not be a continuous numerical sequence under each classification?
Thanks
I have noticed one problem, however.
As an example, in a sort from lowest to highest call numbers, PJ6611 .Z36 2001 precedes PJ6611.B35 2014, which is probably not correct. That is, I think that spaces between classification and Cutter fields should be ignored when sorting as they are not in fact semantic, even though sometimes the first Cutter is actually used as more of an extension of the classification field itself rather than for local shelving purposes. If so, the relative positions of the two call numbers I quote should be reversed.
So, with the same ascending sorting order as previously, I see, for example, call numbers in the following order:
PC4074.7 .P46 2000
PC4075
PC43 .O95 2016
PC45 .P4 1976
PC4075 .P69 2001
(Intermediate items are omitted)
However, if I re-sort (again by Call Number), 'see' becomes 'saw', in that the sorting order has changed overall, again with some examples of mis-sorting in the new arrangement.
There doesn't seem to be any obvious pattern or predictability, at least as far as I can see.
(Zotero is at the 6.0.6 release)