Querying DAU/MAU over time (daily)

Question

I have a daily sessions table with columns user_id and date. I\'d like to graph out DAU/MAU (daily active users / monthly active users) on a daily basis. For example:

Answer 1

You didn't show us your complete table definition, but maybe something like this:

select date,
       count(*) over (partition by date_trunc('day', date) order by date) as dau,
       count(*) over (partition by date_trunc('month', date) order by date) as mau
from sessions
order by date;

To get the percentage without repeating the window functions, just wrap this in a derived table:

select date, 
       dau,
       mau,
       dau::numeric / (case when mau = 0 then null else mau end) as pct
from (
    select date,
           count(*) over (partition by date_trunc('day', date) order by date) as dau,
           count(*) over (partition by date_trunc('month', date) order by date) as mau
    from sessions
) t
order by date;

Here is an example output:

postgres=> select * from sessions;
 session_date | user_id
--------------+---------
 2014-05-01   |       1
 2014-05-01   |       2
 2014-05-01   |       3
 2014-05-02   |       1
 2014-05-02   |       2
 2014-05-02   |       3
 2014-05-02   |       4
 2014-05-02   |       5
 2014-06-01   |       1
 2014-06-01   |       2
 2014-06-01   |       3
 2014-06-02   |       1
 2014-06-02   |       2
 2014-06-02   |       3
 2014-06-02   |       4
 2014-06-03   |       1
 2014-06-03   |       2
 2014-06-03   |       3
 2014-06-03   |       4
 2014-06-03   |       5
(20 rows)

postgres=> select session_date,
postgres->        dau,
postgres->        mau,
postgres->        round(dau::numeric / (case when mau = 0 then null else mau end),2) as pct
postgres-> from (
postgres(>     select session_date,
postgres(>            count(*) over (partition by date_trunc('day', session_date) order by session_date) as dau,
postgres(>            count(*) over (partition by date_trunc('month', session_date) order by session_date) as mau
postgres(>     from sessions
postgres(> ) t
postgres-> order by session_date;
 session_date | dau | mau | pct
--------------+-----+-----+------
 2014-05-01   |   3 |   3 | 1.00
 2014-05-01   |   3 |   3 | 1.00
 2014-05-01   |   3 |   3 | 1.00
 2014-05-02   |   5 |   8 | 0.63
 2014-05-02   |   5 |   8 | 0.63
 2014-05-02   |   5 |   8 | 0.63
 2014-05-02   |   5 |   8 | 0.63
 2014-05-02   |   5 |   8 | 0.63
 2014-06-01   |   3 |   3 | 1.00
 2014-06-01   |   3 |   3 | 1.00
 2014-06-01   |   3 |   3 | 1.00
 2014-06-02   |   4 |   7 | 0.57
 2014-06-02   |   4 |   7 | 0.57
 2014-06-02   |   4 |   7 | 0.57
 2014-06-02   |   4 |   7 | 0.57
 2014-06-03   |   5 |  12 | 0.42
 2014-06-03   |   5 |  12 | 0.42
 2014-06-03   |   5 |  12 | 0.42
 2014-06-03   |   5 |  12 | 0.42
 2014-06-03   |   5 |  12 | 0.42
(20 rows)

postgres=>