Some of the group functions (and probably also of the summary functions) would be more useful if users could select the columns they apply to.
Test case: Buckeye corpus, query string *, data columns POS and Word. Users might want to restrict their output to those POS types where there are at least 50 tokens. They can't use Frequency as a group or summary function for that, because as long as Word is selected, the function will count each combination of POS and Word.
If users could specify that Frequency uses only POS, but ignores Word, a filter could be added that uses the result.
Some of the group functions (and probably also of the summary functions) would be more useful if users could select the columns they apply to.
Test case: Buckeye corpus, query string
*
, data columnsPOS
andWord
. Users might want to restrict their output to those POS types where there are at least 50 tokens. They can't use Frequency as a group or summary function for that, because as long asWord
is selected, the function will count each combination of POS and Word.If users could specify that Frequency uses only
POS
, but ignoresWord
, a filter could be added that uses the result.