Use count_big in aggregate covariates

[anthonysena]

It has been reported that this query:

Characterization/R/AggregateCovariates.R

Lines 260 to 270 in 67613dc

	message("Extracting target cohort counts")
	sql <- "select
	cohort_definition_id,
	count(*) row_count,
	count(distinct subject_id) person_count,
	min(datediff(day, cohort_start_date, cohort_end_date)) min_exposure_time,
	avg(datediff(day, cohort_start_date, cohort_end_date)) mean_exposure_time,
	max(datediff(day, cohort_start_date, cohort_end_date)) max_exposure_time
	from
	(select * from #agg_cohorts_before union select * from #agg_cohorts_extras) temp
	group by cohort_definition_id;"

Is overflowing the int data type. This should use COUNT_BIG instead.