Also if you just don't want the 'name' field included in the super search you can specify that as well in the AnalyzerDefinition by setting fullTextIndex on the ColumnDefinition to false.
~Garrett On Tue, Sep 11, 2012 at 10:41 PM, Tim Williams <[email protected]> wrote: > On Tue, Sep 11, 2012 at 10:23 PM, <[email protected]> wrote: >> My question is about 'super' field in Blur. >> >> Suppose for column family 'customers' we have a column 'name'. >> >> A typical value for such column is a string containing the first and the >> last name >> >> of a person. E.g. 'Adam Johnson'. >> >> If we want to make this field non-analyzable we may specify for it >> >> KeywordAnalyzer instead of StandardAnalyzer. >> >> In this case the search 'customers.name:Adam' will give no results. >> >> It works!. >> >> However search 'Adam' [no qualifiers] returns the record with 'Adam >> Johnson'. >> >> I think that it is done by the 'super' search. >> >> Indeed if we check for all Lucene terms in the shard 'Adam' and 'Johnson' >> are there >> >> related to the field 'super'. >> >> My question is : is it possible to build the table in such a way that the >> values for the field 'name' >> >> would NOT create terms for the 'super' field? >> >> If so, how to specify proper parameters at the table creation? > > Hi Victor, > Yes, an earlier version of Blur respected the Analyzer of the > ColumnDefinition but disregarded the 'fulltext' qualifier and > preferred the default 'fulltext' over a value specified in the > ColumnDefinition. You got it right though, that 'fulltext' equates to > the field "super" underneath the covers. This is fixed in the trunk > and it now behaves just as you suspect. If you can update to the > trunk, it'll "just work" exactly as desired:) > > Thanks, > --tim
