Re: [libreoffice-users] Re: Filter algorithm question

At 11:07 23/02/2015 -0700, Nobody Noname wrote:
> I don't understand the COUNTIF() condition which is why I was 
> looking at the Flood Fill function to search within the sanctuary 
> boundary. If I had a list of X & Y coordinates defining the boundary 
> of the sanctuary I suppose I could search from left to right along 
> the X-axis, then drop down one Y-axis increment searching the inner 
> area and looking for inner coordinates that represented birds. I 
> think I may be talking BASIC and you may be talking LibreOffice.
>
> Whether I filter the X & Y coordinates by reading a serial file in 
> some flavor of BASIC or process them in a spreadsheet like 
> LibreOffice Calc is irrelevant to me, its the algorithm to do it 
> that I'm looking for. However I understand this is a LibreOffice 
> forum so I understand if that's where your coming from. So I 
> probably need to look up the COUNTIF() function.

I assumed (for obvious reasons!) that you would be using a 
spreadsheet, but I think the question of whether you use a program or 
a spreadsheet to solve this problem is a red herring.

You have a list of bird locations. You are proposing to construct a 
separate list of cells that are located within your shape. Then it 
seems that you are going to step through your long list of cells and 
for each cell run through the entire list of bird locations to see 
how many birds are in that cell. That is surely a very inefficient 
way of doing what you ask? You would need something like this only if 
you needed to know separately how many bird locations were positioned 
in *each* cell.

But your declared problem is merely counting the single total number 
of bird locations within your shape. To do this, all you need to do 
is to create a list of the Y values of the shape boundary for each X 
coordinate in your grid (or vice versa). Provided that the shape is 
not re-entrant in at least one direction, you should be able to have 
just two values for each row (or column). So this list has only 
twice-N values instead of the N-squared values in yours. Now you step 
through the *birds* - not the cells - and for each bird you first 
determine the column (or row) in which it occurs and then simply 
whether it is within the limits of the shape in that row in order to 
decide whether to count it. That's far smaller data and far fewer 
calculations - and no need for the "fill" algorithm you mention.

Brian Barker  


--
To unsubscribe e-mail to: users+unsubscribe@global.libreoffice.org
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://listarchives.libreoffice.org/global/users/
All messages sent to this list will be publicly archived and cannot be deleted