tailaijin
commented
8 years ago After adding some index, MBZDB enjoys a much better efficient. I dropped all the tables without any record. Some queries are implemented to insight the data.
`
4. Validate some distribution
NULL year or 2017/2019 must be wrong # Delete in further analysis
SELECT COUNT(re.name), rc.date_year FROM release AS re, release_country AS rc
WHERE rc.release = re.id
GROUP BY (rc.date_year);
Detach unreseasonable date
SELECT COUNT(name), begin_date_year, end_date_year
FROM artist
WHERE begin_date_year > 1950
AND begin_date_year < 2017
AND (end_date_year > 1990 OR end_date_year IS NULL)
GROUP BY(begin_date_year);
Varioulsy types of place
SELECT COUNT(*), place_type.name FROM area, place_type
WHERE area.type = place_type.id
GROUP BY(area.type);
the US has 30 Washington Country...
SELECT COUNT(gid), name FROM area
GROUP BY(name)
ORDER BY(COUNT(gid)) DESC
SELECT * FROM place SELECT * FROM place_type`
The idea now is filtering by time / area/ genre to get a smaller dataset.