I'm having an issue with an old cluster (7.2.1) having some term fields being in uppercase, but a new cluster (7.5.2) having the same terms, from the same docs and synonyms file, being lowercase. Here are the most relevant settings/mappings (I just copied from old to new)
Old cluster:
"filter" : [
"trim",
"lowercase",
"hpo_autophrase_synonyms",
"hpo_entity_extraction_synonyms",
"keep_synonyms"
],
"tokenizer" : "standard"
}
"hpo_terms" : {
"type" : "text",
"term_vector" : "with_positions_offsets",
"eager_global_ordinals" : true,
"analyzer" : "hpo",
"fielddata" : true
}
New cluster:
"filter" : [
"trim",
"lowercase",
"hpo_autophrase_synonyms",
"hpo_entity_extraction_synonyms",
"keep_synonyms"
],
"tokenizer" : "standard"
}
"hpo_terms" : {
"type" : "text",
"term_vector" : "with_positions_offsets",
"eager_global_ordinals" : true,
"analyzer" : "hpo",
"fielddata" : true
}
The old cluster, with the same synonym bundle/plugin as the new, will have terms like "HP_0000001", and then new will have "hp_0000001". I can't seem to find any difference between the settings or mappings.
Any thoughts? Is there some global setting somewhere to force field values to be uppercase?