Conversation
Instead of keeping the XML file originating from Solr 7 around, let's put our customisations in the vanilla config. The indentation changed, so this looks like an "everything changed" thing. Aside from that, we are doing three changes: 1. Enable ClassicIndexSchemaFactory 2. Disable schemaless via "update.autoCreateFields:false" 3. Tweak our searches in /select request handler
- No longer use Trie*Fields, switch all to point fields (Trie has been deprecated since 9) - Add some new field types added since Solr 9 came out
|
I just did a couple "diffs" of vanilla Solr 9.3.0 vs the versions in this pull request. The diff for solrconfig.xml makes perfect sense to me. Only a few lines changed: solrconfigdiff.txt 🚀 I'm find it hard to understand why the diff is so much bigger for schema.xml: schemadiff.txt . It's much harder to reason about. 🤔 It makes me wonder if a script was used to create schema.xml? If so, where is the script? How would I reproduce these results? @poikilotherm I love the approach but let's discuss! Thanks for the PR! |
| <field name="publicationStatus" type="string" stored="true" indexed="true" multiValued="true"/> | ||
| <field name="externalStatus" type="string" stored="true" indexed="true" multiValued="false"/> | ||
| <field name="embargoEndDate" type="long" stored="true" indexed="true" multiValued="false"/> | ||
| <field name="embargoEndDate" type="plong" stored="true" indexed="true" multiValued="false"/> |
There was a problem hiding this comment.
@qqmyers do you remember why you picked long for embargoEndDate when releasedate uses int?
There was a problem hiding this comment.
Not really. I see in the IndexServiceBean that the embargo date is turned into a long via toEpochDay() where the release date is indexed directly as a timestamp, but I'm not sure why that is.
3 participants
Replacing with files compiled from using the vanilla 9.3 config / schema and applying our tweaks.
This is related to PR #9787 as this PR targets the base branch of that one.