« first day (1452 days earlier)      last day (666 days later) » 

11:54 AM
I tried to check what percentage of posts are spam/rude - probably in a clumsy way.
The queries I have so far: per month and per year.
Is there an easy way to get rid of the "gaps" in the graph? I.e., I want to return percentage as zero if there were no spam posts in the month/year.
I could use FROM spam_cte s LEFT OUTER JOIN counts_cte c ON s.CYear=c.CYear - I do not get zeroes, as I wanted; but at least the graph bi longer has gaps.
 
 
4 hours later…
3:59 PM
@Martin one way to do that is to generate a table with the dates (months, years) you want to report over and then do left outer join with your actual statistics query.
 
4:15 PM
probably the simplest that can work: data.stackexchange.com/meta.mathoverflow/query/1618244
 
The list of dates that I want is available already - in counts_cte.
 
That still seems rather similar to what I had.
In the sense that the thing which I wasn't able to do was to get zeroes in place of "empty" records.
 
Yep, that is what I added. I thought that was what you were struggling with?
 
But maybe having the graph with isolated points is better anyway.
 
@rene Following the link from this message, I got this:
 
4:37 PM
probably only selected the link but didn't press ctrl+c ...
 
Thanks for this - I'll look into that.
I'll have to go AFK soon probably - some stuff IRL.
 
take care. I'm around for another 3 hours or so.
 
5:34 PM
I have included a similar query (by year - rather than by month) into the related posts on MathOverflow Meta: To which extent should spam posts on meta be considered a problem?
 

« first day (1452 days earlier)      last day (666 days later) »