[OPEN-ILS-DEV] Authorities sorting

Elfstrand, Stephen F stephen.elfstrand at mnsu.edu
Mon Jul 1 09:51:15 EDT 2013 


Re the sort.  OCLC sorts in the way you describe? That is odd because

150 $aJazz  $zFrance   $zParis   $y1921-1930.    Is a subset of

150 $aJazz $y1921-1930


The sort order that is the most logical at least to me is

150 $aJazz

150 $aJazz $y1921-1930

150 $aJazz  $y1921-1930 $zFrance   $zParis

150 $aJazz $y1921-1930 $zNorth Carolina $zGastonia.

150 $aJazz  $y1931-1940

150 $aJazz  $y1931-1940  $zFrance   $zParis  .

150 $aJazz $y1931-1940  $zNorth Carolina $zGastonia.

150 $aJazz  $y1941-1950

150 $aJazz  $y1941-1950  $zEngland.

150 $aJazz  $y1941-1950  $zFrance   $zParis  .

Stephen Elfstrand
PALS

 [Message: 1
Date: Mon, 24 Jun 2013 17:16:43 -0400
From: Yamil Suarez <ysuarez at berklee.edu<mailto:ysuarez at berklee.edu>>
Subject: [OPEN-ILS-DEV] What are authorities sorted in the cataloging
        "manage authorities" view?
To: Development Evergreen list
        <open-ils-dev at list.georgialibraries.org<mailto:open-ils-dev at list.georgialibraries.org>>
Message-ID: <40939A6B-4605-47B9-97D1-57A97CA5179B at berklee.edu<mailto:40939A6B-4605-47B9-97D1-57A97CA5179B at berklee.edu>>
Content-Type: text/plain; charset=windows-1252

Hello,

My catalogers are confused by how authorities are sorted in the cataloging "manage authorities" view. I wanted to understand the approach being taken by the current code, versus what the catalogers are expecting.


For example in our EG 2.2 and our now EG 2.4 system if we search for subject "jazz" the results look something like this (though the sub-field letters are not displayed)?


150 $aJazz

150 $aJazz $y1921-1930

150 $aJazz  $zFrance   $zParis   $y1921-1930.

150 $aJazz $zNorth Carolina $zGastonia $y1921-1930.

150 $aJazz  $y1931-1940

150 $aJazz  $zFrance   $zParis   $y1931-1940.

150 $aJazz $zNorth Carolina $zGastonia $y1931-1940.

150 $aJazz  $y1941-1950

150 $aJazz  $zEngland $y1941-1950.

150 $aJazz  $zFrance   $zParis   $y1941-1950.


The catalogers would prefer that the sorting should instead look like this?.

150 $aJazz

150 $aJazz $y1921-1930

150 $aJazz  $y1931-1940

150 $aJazz  $y1941-1950

150 $aJazz  $zFrance   $zParis   $y1921-1930.

150 $aJazz  $zFrance   $zParis   $y1931-1940.

150 $aJazz  $zFrance   $zParis   $y1941-1950.

150 $aJazz $zNorth Carolina $zGastonia $y1921-1930.

150 $aJazz $zNorth Carolina $zGastonia $y1931-1940.

150 $aJazz  $zEngland $y1941-1950.


Which is the way that the authorities searches look like within OCLC Connexion client when searching the OCLC authority file. I wonder if part of the confusion is that sometimes authority subfield letters are not sued in alphabetical order. For example, in these subject authority the subfield are used in this order $a, $z, $y.

My understanding is that this old LP bug is referring to this sorting issue.

https://bugs.launchpad.net/evergreen/+bug/781008

This old bug needs to be re-filed because it combined more than one issue at a time, and I am just trying to research authority sorting to see if it should be refiled or turned into a wish-list item, etc.

Thanks in advance,
Yamil










------------------------------

Message: 2
Date: Mon, 24 Jun 2013 17:57:06 -0400
From: Dan Scott <dan at coffeecode.net<mailto:dan at coffeecode.net>>
Subject: Re: [OPEN-ILS-DEV] What are authorities sorted in the
        cataloging "manage authorities" view?
To: Evergreen Development Discussion List
        <open-ils-dev at list.georgialibraries.org<mailto:open-ils-dev at list.georgialibraries.org>>
Message-ID: <20130624215704.GA24800 at denials.eastlink.ca<mailto:20130624215704.GA24800 at denials.eastlink.ca>>
Content-Type: text/plain; charset=us-ascii

On Mon, Jun 24, 2013 at 05:16:43PM -0400, Yamil Suarez wrote:
> Hello,
>
> My catalogers are confused by how authorities are sorted in the
> cataloging "manage authorities" view. I wanted to understand the
> approach being taken by the current code, versus what the catalogers
> are expecting.

The code in authority_tag_sf_browse() in OpenILS/Application/SuperCat.pm is sorting purely by the value of the subfield that was being searched; other subfields currently play no actual role in the sorting. (That is, if the code hasn't changed significantly over the past couple of years).

In addition, the examples you gave are great, but they're missing the 4xx / 5xx fields which can also play a role in the sorting, as you might get a match for a 4xx "see from" or 5xx "see also" field, per http://www.loc.gov/marc/authority/ad400.html.

For example, if you searched for "Singh, Bhagat", you might be surprised to see "Bhagata Si" appearing apparently out of order. However, the idea was that if you displayed the pertinent fields for the hit, you would see that it was a "see from" reference and you would select the authorized result. Obviously there's lots of improvement that could go into the UI for this...


------------------------------

Message: 3
Date: Tue, 25 Jun 2013 09:21:28 -0400
From: Mike Rylander <mrylander at gmail.com<mailto:mrylander at gmail.com>>
Subject: Re: [OPEN-ILS-DEV] What are authorities sorted in the
        cataloging "manage authorities" view?
To: Evergreen Development Discussion List
        <open-ils-dev at list.georgialibraries.org<mailto:open-ils-dev at list.georgialibraries.org>>
Message-ID:
        <CAO8ar=kCad-qLMUVskMc32kUcNZuTsnHR3j6Yvd18kywVbU3Tw at mail.gmail.com<mailto:CAO8ar=kCad-qLMUVskMc32kUcNZuTsnHR3j6Yvd18kywVbU3Tw at mail.gmail.com>>
Content-Type: text/plain; charset=windows-1252

The primary cause of the sort order you're seeing is the code at lines
277 through 279 of Open-ILS/src/sql/Pg/011.schema.authority.sql.

For each tag that we're going turn into simple heading, primarily for browsing authority records, we look only at the subfields that we've been told to.  However, instead of looping over the subfields in record-order, we pull them out of the tag in configuration order.
IOW, record order is not preserved for sorting.

I think this needs a new bug, since the old one you pointed out has a muddled history, and doesn't call out the problem specifically.

The fix for this is relatively straight forward, and I'll see if I can put that together soon.

One last thing, however, regarding your cataloger's desired sort order; it would look like the following (note the placement of the "Jazz -- England" line in your desired example):

150 $aJazz

150 $aJazz $y1921-1930

150 $aJazz  $y1931-1940

150 $aJazz  $y1941-1950

150 $aJazz  $zEngland $y1941-1950.

150 $aJazz  $zFrance   $zParis   $y1921-1930.

150 $aJazz  $zFrance   $zParis   $y1931-1940.

150 $aJazz  $zFrance   $zParis   $y1941-1950.

150 $aJazz $zNorth Carolina $zGastonia $y1921-1930.

150 $aJazz $zNorth Carolina $zGastonia $y1931-1940.


On Mon, Jun 24, 2013 at 5:16 PM, Yamil Suarez <ysuarez at berklee.edu<mailto:ysuarez at berklee.edu>> wrote:
> Hello,
>
> My catalogers are confused by how authorities are sorted in the cataloging "manage authorities" view. I wanted to understand the approach being taken by the current code, versus what the catalogers are expecting.
>
>
> For example in our EG 2.2 and our now EG 2.4 system if we search for subject "jazz" the results look something like this (though the sub-field letters are not displayed)?
>
>
> 150 $aJazz
>
> 150 $aJazz $y1921-1930
>
> 150 $aJazz  $zFrance   $zParis   $y1921-1930.
>
> 150 $aJazz $zNorth Carolina $zGastonia $y1921-1930.
>
> 150 $aJazz  $y1931-1940
>
> 150 $aJazz  $zFrance   $zParis   $y1931-1940.
>
> 150 $aJazz $zNorth Carolina $zGastonia $y1931-1940.
>
> 150 $aJazz  $y1941-1950
>
> 150 $aJazz  $zEngland $y1941-1950.
>
> 150 $aJazz  $zFrance   $zParis   $y1941-1950.
>
>
> The catalogers would prefer that the sorting should instead look like this?.
>
> 150 $aJazz
>
> 150 $aJazz $y1921-1930
>
> 150 $aJazz  $y1931-1940
>
> 150 $aJazz  $y1941-1950
>
> 150 $aJazz  $zFrance   $zParis   $y1921-1930.
>
> 150 $aJazz  $zFrance   $zParis   $y1931-1940.
>
> 150 $aJazz  $zFrance   $zParis   $y1941-1950.
>
> 150 $aJazz $zNorth Carolina $zGastonia $y1921-1930.
>
> 150 $aJazz $zNorth Carolina $zGastonia $y1931-1940.
>
> 150 $aJazz  $zEngland $y1941-1950.
>
>
> Which is the way that the authorities searches look like within OCLC Connexion client when searching the OCLC authority file. I wonder if part of the confusion is that sometimes authority subfield letters are not sued in alphabetical order. For example, in these subject authority the subfield are used in this order $a, $z, $y.
>
> My understanding is that this old LP bug is referring to this sorting issue.
>
> https://bugs.launchpad.net/evergreen/+bug/781008
>
> This old bug needs to be re-filed because it combined more than one issue at a time, and I am just trying to research authority sorting to see if it should be refiled or turned into a wish-list item, etc.
>
> Thanks in advance,
> Yamil
>
>
>
>
>
>
>
>



--
Mike Rylander
 | Director of Research and Development
 | Equinox Software, Inc. / Your Library's Guide to Open Source  | phone:  1-877-OPEN-ILS (673-6457)  | email:  miker at esilibrary.com<mailto:miker at esilibrary.com>  | web:  http://www.esilibrary.com


------------------------------

Message: 4
Date: Tue, 25 Jun 2013 10:13:32 -0400
From: Mike Rylander <mrylander at gmail.com<mailto:mrylander at gmail.com>>
Subject: Re: [OPEN-ILS-DEV] What are authorities sorted in the
        cataloging "manage authorities" view?
To: Evergreen Development Discussion List
        <open-ils-dev at list.georgialibraries.org<mailto:open-ils-dev at list.georgialibraries.org>>
Message-ID:
        <CAO8ar=kZ1-NwaMkQYRSKS8Wq9B2OrAxTVRswXu9EhYhPN3JsZw at mail.gmail.com<mailto:CAO8ar=kZ1-NwaMkQYRSKS8Wq9B2OrAxTVRswXu9EhYhPN3JsZw at mail.gmail.com>>
Content-Type: text/plain; charset=windows-1252

Actually, on second thought, I think that bug is fine for this.  I'll
just use that.  See:
http://git.evergreen-ils.org/?p=working/Evergreen.git;a=shortlog;h=refs/heads/collab/miker/authority-sf-file-order

--miker


On Tue, Jun 25, 2013 at 9:21 AM, Mike Rylander <mrylander at gmail.com<mailto:mrylander at gmail.com>> wrote:
> The primary cause of the sort order you're seeing is the code at lines
> 277 through 279 of Open-ILS/src/sql/Pg/011.schema.authority.sql.
>
> For each tag that we're going turn into simple heading, primarily for
> browsing authority records, we look only at the subfields that we've
> been told to.  However, instead of looping over the subfields in
> record-order, we pull them out of the tag in configuration order.
> IOW, record order is not preserved for sorting.
>
> I think this needs a new bug, since the old one you pointed out has a
> muddled history, and doesn't call out the problem specifically.
>
> The fix for this is relatively straight forward, and I'll see if I can
> put that together soon.
>
> One last thing, however, regarding your cataloger's desired sort
> order; it would look like the following (note the placement of the
> "Jazz -- England" line in your desired example):
>
> 150 $aJazz
>
> 150 $aJazz $y1921-1930
>
> 150 $aJazz  $y1931-1940
>
> 150 $aJazz  $y1941-1950
>
> 150 $aJazz  $zEngland $y1941-1950.
>
> 150 $aJazz  $zFrance   $zParis   $y1921-1930.
>
> 150 $aJazz  $zFrance   $zParis   $y1931-1940.
>
> 150 $aJazz  $zFrance   $zParis   $y1941-1950.
>
> 150 $aJazz $zNorth Carolina $zGastonia $y1921-1930.
>
> 150 $aJazz $zNorth Carolina $zGastonia $y1931-1940.
>
>
> On Mon, Jun 24, 2013 at 5:16 PM, Yamil Suarez <ysuarez at berklee.edu<mailto:ysuarez at berklee.edu>> wrote:
>> Hello,
>>
>> My catalogers are confused by how authorities are sorted in the cataloging "manage authorities" view. I wanted to understand the approach being taken by the current code, versus what the catalogers are expecting.
>>
>>
>> For example in our EG 2.2 and our now EG 2.4 system if we search for subject "jazz" the results look something like this (though the sub-field letters are not displayed)?
>>
>>
>> 150 $aJazz
>>
>> 150 $aJazz $y1921-1930
>>
>> 150 $aJazz  $zFrance   $zParis   $y1921-1930.
>>
>> 150 $aJazz $zNorth Carolina $zGastonia $y1921-1930.
>>
>> 150 $aJazz  $y1931-1940
>>
>> 150 $aJazz  $zFrance   $zParis   $y1931-1940.
>>
>> 150 $aJazz $zNorth Carolina $zGastonia $y1931-1940.
>>
>> 150 $aJazz  $y1941-1950
>>
>> 150 $aJazz  $zEngland $y1941-1950.
>>
>> 150 $aJazz  $zFrance   $zParis   $y1941-1950.
>>
>>
>> The catalogers would prefer that the sorting should instead look like this?.
>>
>> 150 $aJazz
>>
>> 150 $aJazz $y1921-1930
>>
>> 150 $aJazz  $y1931-1940
>>
>> 150 $aJazz  $y1941-1950
>>
>> 150 $aJazz  $zFrance   $zParis   $y1921-1930.
>>
>> 150 $aJazz  $zFrance   $zParis   $y1931-1940.
>>
>> 150 $aJazz  $zFrance   $zParis   $y1941-1950.
>>
>> 150 $aJazz $zNorth Carolina $zGastonia $y1921-1930.
>>
>> 150 $aJazz $zNorth Carolina $zGastonia $y1931-1940.
>>
>> 150 $aJazz  $zEngland $y1941-1950.
>>
>>
>> Which is the way that the authorities searches look like within OCLC Connexion client when searching the OCLC authority file. I wonder if part of the confusion is that sometimes authority subfield letters are not sued in alphabetical order. For example, in these subject authority the subfield are used in this order $a, $z, $y.
>>
>> My understanding is that this old LP bug is referring to this sorting issue.
>>
>> https://bugs.launchpad.net/evergreen/+bug/781008
>>
>> This old bug needs to be re-filed because it combined more than one issue at a time, and I am just trying to research authority sorting to see if it should be refiled or turned into a wish-list item, etc.
>>
>> Thanks in advance,
>> Yamil
>>
>>
>>
>>
>>
>>
>>
>>
>
>
>
> --
> Mike Rylander
>  | Director of Research and Development
>  | Equinox Software, Inc. / Your Library's Guide to Open Source
>  | phone:  1-877-OPEN-ILS (673-6457)
>  | email:  miker at esilibrary.com<mailto:miker at esilibrary.com>
>  | web:  http://www.esilibrary.com



--
Mike Rylander
 | Director of Research and Development
 | Equinox Software, Inc. / Your Library's Guide to Open Source
 | phone:  1-877-OPEN-ILS (673-6457)
 | email:  miker at esilibrary.com<mailto:miker at esilibrary.com>
 | web:  http://www.esilibrary.com


End of Open-ils-dev Digest, Vol 87, Issue 12
********************************************

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://libmail.georgialibraries.org/pipermail/open-ils-dev/attachments/20130701/2b7703ca/attachment-0001.htm>

More information about the Open-ils-dev mailing list