[Archivesspace_Users_Group] Archon to ArchivesSpace Migration Issues

[Karen Miller]

Hi, Suzanne.

We migrated from Archon to ArchivesSpace 1.4.2 in December at Northwestern University. We did not experience issues with top containers, but we did have about 2 dozen (out of 1200+) finding aids come over with scrambled components.

As you described, the scrambling appeared to be completely random. When I looked at records in the Archon MySQL database for a few of the scrambled components, I found some components with parent relationships that formed a circle - for example, the parent ID values of records indicated that component A had component B as parent, component B had component C as parent, and component C had component A as parent. We think that was due to people moving around the components in the Archon interface, or that it might have had to do with a flawed ingest into Archon from the original EAD. Some of these components did not appear in the Archon interface, only in the MySQL database.

For the "scrambled" finding aids, I exported these as EAD from Archon and then used the Background Job -> Import Data to ingest them into ArchivesSpace. This is not without its own drama. Ingesting from EAD this way has a few bugs:


·        Agent & Subject records will come in as Unpublished (this might be rectified if you set the Publish? checkbox defaults to be checked in User Preferences, but I'm not sure about that).

·        Ampersands display in ArchivesSpace as character entities (&) instead of as ampersands.

·        Any <note> elements attached to a component (Archival Object) record did not ingest at all. <note> elements at the Resource record did ingest, so this is only a problem for really detailed finding aids.

To make the ingest work, you'll need to add the <extent> type attribute into the value for the <extent> element at the archival description level. That is, under <archdesc>, change the line that looks like this:

<physdesc label="Physical Description">
<extent encodinganalog="300" type="Boxes">121</extent>

to this:

<physdesc label="Physical Description">
<extent encodinganalog="300" type="Boxes">121 Boxes</extent>

You'll also want to change the value of the <eadid> value to something that differentiates it from the <eadid> value for the existing resource record in ArchivesSpace. I added "-TEST" to the end of mine so that I could differentiate them in ArchivesSpace.

Finally, you might run into errors because of components that lack a <unittitle> and a <unitdate> during the ingest. You can fix these and then reingest the finding aid, but this can get a little tedious because the ingest only identifies the first one it runs into and then fails. (So if you have 5 components that cause it errors, you'll have to try the ingest 6 times to get it in there). Interpreting the errors is tedious, but it can be done. They will look like this:

The following errors were found:
        dates : one or more required (or enter a Title)
        title : must not be an empty string (or enter a Date)


For JSONModel(:archival_object):
 #<JSONModel(:archival_object) {"jsonmodel_type"=>"archival_object", "external_ids"=>[], "subjects"=>[], "linked_events"=>[], "extents"=>[], "dates"=>[], "external_documents"=>[], "rights_statements"=>[], "linked_agents"=>[], "component_id"=>"id113570", "restrictions_apply"=>false, "instances"=>[{"jsonmodel_type"=>"instance", "is_representative"=>false, "uri"=>nil, "import_context"=>"<container class=\"cdata\" type=\"Box\"> ... </container>", "instance_type"=>"mixed_materials", "container"=>{"jsonmodel_type"=>"container", "container_locations"=>[], "uri"=>nil, "import_context"=>"<container class=\"cdata\" type=\"Box\"> ... </container>", "container_profile_key"=>nil, "type_1"=>"Box", "indicator_1"=>"89", "type_2"=>"Folder", "indicator_2"=>"1", "type_3"=>"Sleeve", "indicator_3"=>"24"}}], "notes"=>[], "uri"=>"/repositories/import/archival_objects/import_544f0225-8313-4396-9c17-0ed8075640fc", "level"=>"file", "resource"=>{"ref"=>"/repositories/import/resources/import_73c30f1d-c9e2-444b-b482-84a380ee8f1a"}, "parent"=>{"ref"=>"/repositories/import/archival_objects/import_be1e2fd1-38d2-4550-be86-efdd88070162"}, "publish"=>true}>


In :
 <c03 audience="external" class="cdata" level="file"> ... </c03>


!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

This one means that a component with container elements representing Box 89, Folder 1, and Sleeve 24 (bolding & font color above are mine and I hope they come through the listserv) has neither a <unittitle> nor a <unitdate>.

I hope this is helpful. This worked really well for our smaller finding aids, but it wasn't ideal for a handful of large ones for us. Still, it was better than trying to unscramble the components in ArchivesSpace.

Karen


Karen D. Miller
Monographic Cataloger/Metadata Specialist
Northwestern University Libraries
Northwestern University
1970 Campus Drive
Evanston, IL 60208
www.library.northwestern.edu
k-miller3 at northwestern.edu
874.467.3462

From: archivesspace_users_group-bounces at lyralists.lyrasis.org [mailto:archivesspace_users_group-bounces at lyralists.lyrasis.org] On Behalf Of Stasiulatis, Suzanne
Sent: Wednesday, March 08, 2017 7:43 AM
To: archivesspace_users_group at lyralists.lyrasis.org
Subject: [Archivesspace_Users_Group] Archon to ArchivesSpace Migration Issues

Hello,

We have two major problems with our migration from Archon to ArchivesSpace. Any help is appreciated. We did not have success with the migration from Archon to 1.4.2. We moved to 1.0.4 to 1.4.2 to 1.5.2. The migration and upgrade complete but we have two major problems.

There are no top container drop downs in the component tree. The top containers are there, but they are not appearing as expected with drop downs in the component tree or in the staff interface.

A large portion of our component data is scrambled. Some of the components are perfectly fine. They are in order and match the Archon listing line for line. Other listings (about 1/3 to ¼) are out of order. Often the components are in order within a container, but out of order otherwise. Other listings have components out of order. We've searched for a pattern but not found one. The scrambling seems to be totally random.

Thanks in advance!

Suzanne

Suzanne Stasiulatis | Archivist II
Pennsylvania Historical and Museum Commission | Pennsylvania State Archives
350 North Street | Harrisburg, PA 17120-0090
Phone: 717-787-5953
http://www.phmc.pa.gov<https://urldefense.proofpoint.com/v2/url?u=http-3A__www.phmc.pa.gov_&d=CwMFAw&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=P54V6LrXJP5Ddzc8ZItBdqB1Kr3elvFIQ04P7n0UCbQ&m=6l7pcX_DRoO3WDrM3UVIW_dUQy5XDnwCyQqCWS6brZ4&s=37d1H2sjeBYbYn-9kBvzcgmiwkexfB5gqsAWqtLQBEw&e=>
sustasiula at pa.gov<mailto:sustasiula at pa.gov>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lyralists.lyrasis.org/pipermail/archivesspace_users_group/attachments/20170308/10676721/attachment.html>