Measure queries

Here is a table of the kinds of queries that we make when calculating measures.

count	columns	from
Analyse-style queries against prescribing
50	SUM(items)	{hscic}.normalised_prescribing_standard
18	SUM(actual_cost)	{hscic}.normalised_prescribing_standard
16	SUM(quantity)	{hscic}.normalised_prescribing_standard
1	SUM(denominator)/17	{measures}.practice_data_all_low_priority
1	SUM(numerator)	{measures}.practice_data_all_low_priority
Analyse-style queries against practice statistics
24	SUM(total_list_size / 1000.0)	{hscic}.practice_statistics
1	CAST(JSON_EXTRACT(MAX(star_pu), '$.oral_antibacterials_item') AS FLOAT64)	{hscic}.practice_statistics
Custom weightings
1	SUM(lyrica_mg)	{measures}.pregabalin_total_mg
1	SUM(total_ome)	{measures}.opioid_total_ome
1	SUM(p.quantity * r.percent_of_adq)	{hscic}.normalised_prescribing_standard p LEFT JOIN {hscic}.presentation r ON p.bnf_code = r.bnf_code
1	SUM(quantity/ CASE WHEN bnf_name LIKE '%Oral%' THEN 10 WHEN bnf_name LIKE '%Liq%' THEN 10 WHEN RTRIM(bnf_name) LIKE '%10mg' THEN 2 WHEN RTRIM(bnf_name) LIKE '%40mg' THEN 0.5 ELSE 1 END)	{hscic}.normalised_prescribing_standard
Other
1	(max(female_35_44) + max(female_45_54) + max(female_55_64) + max(female_65_74)) / 1000.0	{hscic}.practice_statistics

Generated with the following (and a little bit of tidying up):

nc = Counter(
        (m.numerator_columns.replace('AS numerator,', '').strip(), m.numerator_from.strip())
        for m in Measure.objects.all()
    )
dc = Counter(
        (m.denominator_columns.replace('AS denominator,', '').strip(), m.denominator_from.strip())
        for m in Measure.objects.all()
    )

for (columns, from_clause), count in (nc + dc).most_common():
    print '|', count, '|', columns, '|', from_clause