Hello, wondering if anyone has experienced this when using partial_query on the suggest service if I search for an article with a number in the title it will not return any results.
For example I have indexed items with the titles:
‘2020 Vision’ , ‘Pre-Renewal Report - December 2016’
If I use suggest or a normal search with partial query enabled it will not return any results (normal full query matches work fine) so searching for ‘2020’ or even ‘vision’ doesn’t not return the expected suggestion of the item listed above. This also seems to be the same when looking for matches on meta data classes which contains numbers.
Is there a setting needing to be set in the collection for this to work or could it be an indexing config issue on the collection?
I’ve done some digging and it looks like suggestions that begin with a number are not recorded for the spelling-based suggestions and this appears to be because the spelling suggestion system probably doesn’t record any spelling suggestions that begin with a number.
I have raised this with the developers but in the short term there is probably not much you can do about this. The CSV based suggestions do seem to respect triggers that start with a number. As a workaround you could possibly generate a CSV style auto-completion source and use this in addition to or instead of the spelling based suggestions or you might be able to create a spelling whitelist that has some of the suggestions that begin with a number.