MySQL group by date smaller then multiple dates - mysql

Question

this is a hard one(i think).
i need to group a table by multiple dates . the tricky part is that i dont need the group to be the specific date , i need it to be all values that smaller then the date. so if it was only for one date my query will looks like this.
select count(*) from users where created_at <= '2016-01-01'
so my current and rely bad solution is to just do as many query's as many dates that i have. for example if i have and array of 10 dates that mean my code will look like this
for(let i = 0; i< dates.length;i++){
db.query(`select count(*) from users where created_at <= '${dates[i]}'`)
}
desired output:
dates = [ '2016-06-01' ,' 2016-04-01' ,'2016-02-01']
table users:
id | created_at
1 | 2016-01-01
2 | 2016-02-01
3 | 2016-03-01
4 | 2016-04-01
5 | 2016-05-01
6 | 2016-06-01
outpout:
date | count
'2016-06-01' | 6
'2016-04-01' | 4
'2016-02-01' | 2
is there a way to achieve this with 1 query?

Answer 1

//explained logic using java but hope this will help:)
you can achieve this using union,
first concat all dates and make query using union,
String sqlPart="";
String union="";
for(int i = 0; i< dates.length;i++){
sqlPart=sqlPart+union;
sqlPart=sqlPart + "SELECT created_at , count(*) FROM users WHERE created_at <= "+ dates[i];
union="union";
}
so your final query will form in 'sqlPart'
then run this query

Answer 2

There is way to get it in one database hit, yes some logical things need to do:
Dynamically generate query as like below :
query = "SELECT "
for(let i = 0; i< dates.length;i++){
if(i>0) query +=",";
query += "SUM(if(created_at <='${dates[i]}',1,0))";
}
query += " from users ";
if you know max date in advance then add below part also to get it faster :
query += " WHERE created_at<= 'Maxdate' ";
And finally hit database.
db.query(query);
Here you will get the result like this :
6,4,2
Now here each column value will give the count for date in same sequence of array.
Example :
First value 6 is count for 1st array value which is 2016-06-01 and so on, which can be formatted in any format base on requirement.
Still you can improve this better by applying your logical things.
This is not exact compiled code, take it as reference to achieve the same.

Answer 3

Check this please:
SELECT created_at , count(*)
FROM users
WHERE created_at < '2016-01-01'
GROUP BY created_at;
Query will return grouped dates with count of rows for each date.