Question

i want to split a date list in hours and do operations like an average.
I’ve written a python program that call a sqlite3 database, the query returns a list:

def SQLQueryDaily(currency,start,end):
    #year = start[0,3]
    c.execute('SELECT buy, sell FROM '+currency+' WHERE (datetime > "'+start+'" AND datetime < "'+end+'")')
    for row in c:
        print (row)

and it prints:

(‘2002-01-02 01:33:57’, 0.894)

(‘2002-01-02 01:33:58’, 0.895)

(‘2002-01-02 01:33:59’, 0.893)

and so on for thousands of lines…

what i want to do is to regroup this list into hours and do the average in the number returned (here : 0.894)

I honestly tried to find a way to regroup results by hour or days but i don’t know if there is a proper way to do it, please help thanks

Answer 1

This query can be performed in sqlite. To group by hour:

def SQLQueryDaily(currency,start,end):
    sql = '''
        SELECT buy, AVG(sell)
        FROM {t}
        WHERE (datetime > ? AND datetime < ?)
        GROUP BY strftime('%Y-%m-%d %H',date)
        '''.format(t = currency)
    c.execute(sql, [start, end])
    for row in c:
        print (row)    

To group by 15 minutes:

sql = '''
    SELECT buy, AVG(sell)
    FROM {t}
    WHERE (datetime > ? AND datetime < ?)
    GROUP BY strftime('%s', date)/(15*60)
    '''.format(t = currency)

To get the first and last row of each group:

sql = '''
    SELECT f.date, f.sell, t.minp, t.maxp, t.avgp
    FROM {t} f
    INNER JOIN (
        SELECT MIN(date) mindate, MAX(date) maxdate,
               MIN(sell) minp, MAX(sell) maxp, AVG(sell) avgp
        FROM {t}
        GROUP BY strftime('%s', date)/(15*60)
        ) t
        ON f.date = t.mindate or f.date = t.maxdate
    '''.format(t = currency)