如何使该统计查询动态化？

Question

I have this user table:

id(int) | created (datetime)

I'm showing statistics on a server which does following: - You choose a date - You choose if you want to show daily, weekly or monthly statistics (see query)

I have made the query very static by just doing it like this:

WEEK(DATE_ADD(" + date + ", INTERVAL -2 WEEK) // Shows how many users that has been created from two weeks before current date to current date

Now i want the admin that should look at the statistic to be able to choose how far back in time he wants to show statistics from. So i need to make the query dynamic.

Here is the original static query i have made:

String sQuery = "SELECT " +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created) as totalNumberOfCreatedUsers, " +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE DAYOFYEAR(" + date + ") = DAYOFYEAR(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created)) as thisDaysNumberOfCreatedUsers, " +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE DAYOFYEAR(DATE_ADD(" + date + ", INTERVAL -1 DAY)) = DAYOFYEAR(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created)) as thisDaysMinusOneNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE DAYOFYEAR(DATE_ADD(" + date + ", INTERVAL -2 DAY)) = DAYOFYEAR(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created)) as thisDaysMinusTwoNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE DAYOFYEAR(DATE_ADD(" + date + ", INTERVAL -3 DAY)) = DAYOFYEAR(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created)) as thisDaysMinusThreeNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE DAYOFYEAR(DATE_ADD(" + date + ", INTERVAL -4 DAY)) = DAYOFYEAR(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created)) as thisDaysMinusFourNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE DAYOFYEAR(DATE_ADD(" + date + ", INTERVAL -5 DAY)) = DAYOFYEAR(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created)) as thisDaysMinusFiveNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE DAYOFYEAR(DATE_ADD(" + date + ", INTERVAL -6 DAY)) = DAYOFYEAR(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created)) as thisDaysMinusSixNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE DAYOFYEAR(DATE_ADD(" + date + ", INTERVAL -7 DAY)) = DAYOFYEAR(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created)) as thisDaysMinusSevenNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE WEEK(" + date + ") = WEEK(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created))  as thisWeeksNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE WEEK(DATE_ADD(" + date + ", INTERVAL -1 WEEK)) = WEEK(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created))  as thisWeeksMinusOneNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE WEEK(DATE_ADD(" + date + ", INTERVAL -2 WEEK)) = WEEK(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created))  as thisWeeksMinusTwoNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE WEEK(DATE_ADD(" + date + ", INTERVAL -3 WEEK)) = WEEK(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created))  as thisWeeksMinusThreeNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE WEEK(DATE_ADD(" + date + ", INTERVAL -4 WEEK)) = WEEK(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created))  as thisWeeksMinusFourNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE MONTH(" + date + ") = MONTH(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created))  as thisMonthsNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE MONTH(DATE_ADD(" + date + ", INTERVAL -1 MONTH)) = MONTH(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created))  as thisMonthsMinusOneNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE MONTH(DATE_ADD(" + date + ", INTERVAL -2 MONTH)) = MONTH(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created))  as thisMonthsMinusTwoNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE MONTH(DATE_ADD(" + date + ", INTERVAL -3 MONTH)) = MONTH(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created))  as thisMonthsMinusThreeNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE MONTH(DATE_ADD(" + date + ", INTERVAL -4 MONTH)) = MONTH(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created))  as thisMonthsMinusFourNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE MONTH(DATE_ADD(" + date + ", INTERVAL -5 MONTH)) = MONTH(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created))  as thisMonthsMinusFiveNumberOfCreatedUsers," +
                "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE MONTH(DATE_ADD(" + date + ", INTERVAL -6 MONTH)) = MONTH(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created))  as thisMonthsMinusSixNumberOfCreatedUsers" +
                "";

eg: One solution i could do (java): (But this would create a large query, and i want a more efficient way to do it) (btw, this example is not working code, just something i thought of to give you and idea of what a solution could be)

---

The admin chooses to see 20 days/weeks/months

String addToQuery = "";
int daysWeeksMonthsToLookBack = inputFromUser // inputFromUser = 20 in this example
for(int i=0; i<daysWeeksMonthsToLookBack; i++){
  addToQuery += "       (SELECT count(statistic_customer_created.id) FROM statistic_customer_created WHERE DAYOFYEAR(DATE_ADD(" + date + ", INTERVAL -" + i + " DAY)) = DAYOFYEAR(statistic_customer_created.created) AND YEAR(" + date + ") = YEAR(statistic_customer_created.created)) as thisDaysMinusFiveNumberOfCreatedUsers," +

}

Anyone got a suggestion how i can do this more efficiently? Thanks in advance

Answer 1

I would have used a range like

SELECT count(s.id) 
FROM statistic_customer_created s
WHERE s.created >= {date calculated in Java}

By taking the difference of these dates you can get all those created between 1 and 2 days ago.

The count for those created in the last 7 days and those created in the last week should be the same.