Feature #8053
add funding award number to index
0%
Description
Many groups want to track data sets based on the funding award numbers and organizations that funded the work. For example, the Arctic Data Center, BCO-DMO, and R2R all need to report on and search for data based on NSF award numbers. We should add two new multi-valued fields, funding_agency and award_number to the SOLR index so that they can be used for search, display, and faceting. There is a proposal in EML to add structured fields for these, so for details see EML issue https://github.com/NCEAS/eml/issues/266
History
#1 Updated by Matthew Jones over 7 years ago
- Tracker changed from Task to Feature
#2 Updated by Dave Vieglais over 7 years ago
- Project changed from CN Index to Infrastructure
- Category changed from d1_cn_index_common to dataone-cn-index
- Target version set to CCI-2.4.0
- Milestone set to None
#3 Updated by Bryce Mecum over 5 years ago
I'm working on the EML 2.2 indexing stuff and can close this ticket while I do that. To support pre-2.2 docs and 2.2 and onward, I think we should use an XPath of
//dataset/project/funding/text() | //dataset/project/award/awardNumber/text()
I think this is the best route to go because it allows us create a single way to search for funding across older and newer EML docs. This is opposed to making clients do queries like ?q:funding:X OR awardNumber: X. I decided to make the field called funding and define it as a string type with a copyField declaration to fundingText of type text_general.
<field name="funding" type="string" indexed="true" stored="true" multiValued="true" /> <field name="fundingText" type="text_general" indexed="true" stored="false" multiValued="true" />
I decided to do this also to support legacy docs which often stored funding info as free-text strings like "NSF Award Number 123456789".
Do others like this structure and do we also want to support a 'funderName' field for EML2.2 docs?
#4 Updated by Chris Jones over 5 years ago
Bryce - Yes, good strategy. One question: Since EML < 2.2.0 documents have //dataset/project/funding typed as EMLTextType, the call to text() there may give back nothing if it's not a text node being returned. This will be the case for the Arctic Data Center, which had to use //dataset/project/funding/para to be able to associate multiple funding numbers with a single dataset. I wonder how we can handle the sub-element content of this EMLText node. Thoughts?
