Correction. "mark seattle" query doesn't show suggestions since "mark" alone has
some hits.
It is when the same logic is used for a single term query of "seatle" that 3
suggestions of "seattle"
are returned. Do I have to identify the field by using startOffset value?
On 4/7/20 3:46 PM, TK Solr wrote:
I query on multiple field like:
q=city:(mark seattle) name:(mark seattle) phone:(mark seattle)=true
The raw query terms are distributed to all fields because I don't know what
term is intended to for which field.
If I misspell seattle, I get 3 suggestions:
"spellcheck":{
"suggestions":[
"seatle",{
"numFound":1,
"startOffset":29,
"endOffset":35,
"suggestion":["seattle"]},
"seatle",{
"numFound":1,
"startOffset":50,
"endOffset":56,
"suggestion":["seattle"]},
"seatle",{
"numFound":1,
"startOffset":73,
"endOffset":79,
"suggestion":["seattle"]}]}}
(Please disregard exact numbers. It's from more complicated query of the same
nature.)
I think it's showing a correction suggestion for each query field.
Since the phone field keeps a phone number and spelling corrections are not
very useful,
I would like the spellchecker to skip this and similar fields but I don't see
a relevant
parameter in spellchecker's documentation. Is there any way to specify the
fields I am interested or I am not interested?
TK