Hi,
I had been wanting to change events that are unique over month but to no avail. I will give an illustration below:
''month''  ''event''
''1''            ''a''
''1''            ''b''
''1''            ''c''
''2''            ''a''
''2''            ''c''
''2''            ''z''
''2''            ''d''
''2''            ''z''
''3''            ''a''
''3''            ''z''
''3''            ''b''
''3''            ''g''
''3''            ''h''
''3''            ''u''
''3''            ''z''
''3''            ''b''
assuming the index=someIndex
index=somIndex | timechart dc(event) as ''Unique new count'' by month
give you:
''month''   ''Unique new count''
''1''              ''3''
''2''              ''4''
''3''              ''6'' 
The result I want would be:
''month''   ''Unique new count''
''1''              ''3''
''2''              ''2''
''3''              ''3''
Any one can help?
 
					
				
		
For example, you can count on this search sentence.
However, the month needs a year and month.
(your search)
|table month event
|stats min(month) as month by event
|stats count as "Unique new count" by month
Edit: Another example to clarify the question:
1st month I have 10 unique events
2nd month I have 2 new unique events that did not happened in the 1st month 
3rd month I have 5 new unique events that had not happened in 1st and 2nd month combined
4th month I have 7 new unique events that had not happened in 1st to 3rd month
The plot will be like:
1st - 10
2nd - 2
3rd -5
4th - 7 
 
					
				
		
For example, you can count on this search sentence.
However, the month needs a year and month.
(your search)
|table month event
|stats min(month) as month by event
|stats count as "Unique new count" by month
I tried it on my data, but this is what I get.
For the last 3 month. if I use time chart, my value of Jan and Dec are the same, however the Nov are less.
And I think the sequence are not correct. The earlier month should be the super set of all.. and the subsequent months are incremental of the earlier months.
 
					
				
		
I thought that counting the months when the event first appeared could get unique events every month.
Is it different from yours?
I think I put it wrongly in my question. I looking for a more incremental kinda of counting.
For example: 1st month I have 10 unique events, 2nd month I have 2 new unique events that did not happened in the 1st month, 3rd month I have 5 new unique events that had not happened in 1st and 2nd month combine, and 4th month I have 7 new unique events that had not happened in 1st to 3rd month.
The plot will be like:
1st - 10
2nd - 2
3rd -5
4th - 7 
 
					
				
		
You should be able to see and understand the moon that first appeared.
If I do it is such a search sentence.
(your search)
|stats earliest(_time) as time by event
|eval month=strftime(time,"%Y-%m")
|stats count as "Unique new count" by month
Hi, 
Thanks for the reply, I got it to work. I did a mistake to dudup the event.
Thanks again!
