LISP Request - Block Count Inside a Closed Polyline
If the outlines are always rectangular and orthogonal, I can imagine a way to do it. Would that be the case? It would involve getting the bounding box of the rectangle, finding all Blocks, and stepping through that selection, checking whether the X and Y coordinate values of each one's insertion point fall between those of the bounding box corners.
[There are routines out there to find everything contained within a selected Polyline boundary, but I think they would find only things fully within, which it doesn't seem would do what you want, since some of yours with their insertion points inside a boundary extend partly outside it elsewhere.]
Kent Cooper, AIA 
I don't mind at all 
I've found the counting to be a bit slow myself and are already making changes to eliminate to multiple SSGET's needed as this is causing the slow processing I assume.
But looking at the code you posted; well, I don't have my wizard degree yet, so I'll need to take my time to understand completely how it works 
@DannyNL wrote:
...
But looking at the code you posted; well, I don't have my wizard degree yet, so I'll need to take my time to understand completely how it works
Don't get me wrong, I just have a general awareness about the principle and this sub saved in my archive.
HERE is the original link to the routine, HERE is wiki site about Ray casting algorithm.
All the credits to @Lee_Mac and @_gile.
This may overlap with suggestions already made here, but there are routines out there to determine whether a location [XY coordinates] is inside a Polyline of any shape, with various approaches, such as this one. Could you find all Blocks, extract their insertion points, and check those insertion points on their own for whether they're inside, using a routine such as that, rather than working with the Blocks themselves? That particular routine's approach may take some time with a large number of Blocks, but it looks to me like it would be viable, and wouldn't have the problems some suggestions here have [e.g. with Blocks whose insertion points are outside but whose contents extend inside].
Kent Cooper, AIA
@Kent1Cooper wrote:
.... Could you find all Blocks, extract their insertion points, and check those insertion points on their own for whether they're inside, using a routine such as that, rather than working with the Blocks themselves? That particular routine's approach may take some time with a large number of Blocks....
That routine's method is to temporarily Offset the boundary object Through the desired point, and compare the area of the result with that of the selected boundary. Doing that kind of thing narrowed down to your particular sitution comes out like this [lightly tested and without the usual controls yet]:
(defun C:BIB ; = Blocks Inside Boundary
(/ bndry reflen blks n blk)
(setq
bndry (car (entsel "\nBoundary object: "))
reflen (vlax-curve-getDistAtParam bndry (vlax-curve-getEndParam bndry))
blks (ssget "_X" (list '(0 . "INSERT") (cons 410 (getvar 'ctab))))
insiders (ssadd); initially empty
); setq
(repeat (setq n (sslength blks))
(command "_.offset" "_through" bndry
"_none" (cdr (assoc 10 (entget (setq blk (ssname blks (setq n (1- n)))))))
""
); command
(if (< (vlax-curve-getDistAtParam (entlast) (vlax-curve-getEndParam (entlast))) reflen)
(ssadd blk insiders); put it in the selection
); if
(command "_.u"); Undo Offset
); repeat
(princ)
); defun
It took about 15 seconds to do any rectangle in your sample drawing [it shouldn't matter which, since it compares the area of something Offset through the insertion point of every Block in the current space to whatever boundary was selected]. I tried it successfully with some Blocks partially outside but their insertion points inside, and vice versa, and with a Circle as a boundary, and with a more convoluted-shaped Polyline.
It simply puts the ones whose insertion points are inside the boundary into the non-localized "insiders" selection set. It could put them into a list if that's easier to work with. Processing the result for Block count, table input, etc., would be in addition to what this much does.
Kent Cooper, AIA