Boolean operations on multiple selections

[craftybones]

Given a circle C and a series of concentric squares A1, A2 etc, I would like C to cut across A1...An right down the middle and apply a subtraction on all of them.

 

If I select all of them and try to do it wholesale, it applies as if A3-(A2-A1) etc, when what I really want is (C-A1) U (C-A2) etc.

 

I have attached pictures of the setup and what I actually want. When I apply a grouped subtraction, everything gets deleted, obviously.

 

Is there a simple way of doing this?

 

 

Thank you,

 

craftybones

[PixelPest]

When you´re combining the squares before the Boolean subtract the result is looking like this:

If that´s what you´re after.

 

Cheers

P.

[Alfred]

1. Marquee-select all of the objects

2. Click on 'Divide' on the Geometry toolbar

3. Marquee-select the circle

4. Press the Delete key on your keyboard

[PixelPest]

Mmmh - are you sure he wants 18 individual objects? Just curious.

 

Cheers

P.

[Alfred]

Mmmh - are you sure you get 18 individual objects? Just curious.

 

[PixelPest]

You´re right; it´s 48 actually by dividing these 4 Elements selected with the circle on top:

When the squares are on top it´s just 7 BTW:

 

Cheers

P

[Alfred]

When I did it I got 37 objects, which is why I said to marquee-select the circle instead of just clicking on it. There were 33 objects inside the circle, so deleting those together with the enclosing circle left me with the three objects that you see in my later screenshot.

[R C-R]

Using Alfred's method, at step 2, I get 7 objects:

Because I have AD set for the marquee to select anything it touches, for this particular arrangement if I were to marquee select the circle in his step 3, I would delete some of three desired shapes, but it is easy enough to select just the C-shaped object, delete that & then do the same with a marquee select of the remaining three lefthand ones.

 

EDIT: When I first did this, the ellipse (circle) shape was at the bottom of the layers panel & I got 7 shapes from the Divide action. After reading the other replies about getting a different number of shapes, I tried moving the circle to the top & that resulted in 29 shapes! I never did get 37 shapes, regardless of the layer order I tried, so I think there is some kind of bug in the Divide action...

[PixelPest]

There is a difference in the resulting number of objects just by different types of selecting? Are you sure this isn´t a bug?

What ever it takes I still have no clue what kind of shape the user is after.

 

Cheers

P.

[R C-R]

I don't really understand why the layer order makes such a big a difference in the number of objects that result from the divide operation.

 

PixelPest, in the Tools preferences for AD, I have the "Select object when intersects with selection marquee" option selected, so when I drag out a marquee every shape it touches is selected. Without that option enabled, only objects completely enclosed by the marquee are selected.

[PixelPest]

I don't really understand why the layer order makes such a big a difference in the number of objects that result from the divide operation.

Stacking (Z-order) is essential for Boolean operations in SVG-apps - especially for subtracting.

 

Cheers

P.

[MEB]

The number of objects created varies with the layer position and the size of the objects.

There's already a few bugs with boolean operations logged. Hopefully this will be improved/fixed as soon as there's time to revamp/review the boolean operations code.

[Alfred]

I don't really understand why the layer order makes such a big a difference in the number of objects that result from the divide operation.

 

Me too neither, but that's presumably covered by the list of bugs that MEB mentioned.

 

PixelPest, in the Tools preferences for AD, I have the "Select object when intersects with selection marquee" option selected, so when I drag out a marquee every shape it touches is selected. Without that option enabled, only objects completely enclosed by the marquee are selected.

 

Perhaps I should have mentioned that I'm using the default setting, where that option isn't enabled (so marquee-selecting the circle selects only the circle and the objects which are completely contained within it).

[R C-R]

PixelPest, I am only concerning myself with the Divide boolean in AD for this. Subtraction seems to work as expected regarding stacking/z-order, but not Divide.

 

For example, when I selected only three unfilled concentrically arranged square shapes with no other overlapping shape in the canvas & applied Divide, I got 5 shapes because the two inner squares were duplicated. I would expect in this case for the divide operation to have no effect.

[Alfred]

Stacking (Z-order) is essential for Boolean operations in SVG-apps - especially for subtracting.

 

Cheers

P.

 

Stacking order is essential for determining what is subtracted, but (as far as I can see) it shouldn't make any difference to any of the other Boolean operations.

[Alfred]

PixelPest, I am only concerning myself with the Divide boolean in AD for this. Subtraction seems to work as expected regarding stacking/z-order, but not Divide.

 

For example, when I selected only three unfilled concentrically arranged square shapes with no other overlapping shape in the canvas & applied Divide, I got 5 shapes because the two inner squares were duplicated. I would expect in this case for the divide operation to have no effect.

 

I wouldn't expect Divide to have no effect in that scenario, but I think it should combine objects to leave you with three objects instead of five: the smallest square, the shape which you get when you subtract that square from the next larger one, and the shape which you get when you subtract that middle square from the largest one.

[gdenby]

I tried various sizes, and variations on the layer order. Lowest number count of curves after division was around 30. highest, somewhere over 80.

 

The most expedient method I came up with was to draw the nested squares, and the circle. Copy the circle. Select inner square, subtract. Paste circle, select next, subtract. Repeat. Took maybe a minute. Far less painful than trying to sort thru the arc fragments made by dividing.

[R C-R]

Alfred, I expected it to have no effect because the 3 square rectangle objects were unfilled. However, starting with filled square rectangle objects of different colors, I still get 5 objects, each converted a to 4 node curve object. It is more obvious what is happening when I separate them out:

Instead of creating two curves (compound) objects for the larger two, it just duplicates the objects needed to create those curves (compound) objects.

 

gdenby, when I place the circle at the bottom of the layer stack, I get just 7 objects after the Divide action. Did you try that?

Edited October 17, 2016 by R C-R

[Alfred]

I tried various sizes, and variations on the layer order. Lowest number count of curves after division was around 30. highest, somewhere over 80.

 

The most expedient method I came up with was to draw the nested squares, and the circle. Copy the circle. Select inner square, subtract. Paste circle, select next, subtract. Repeat. Took maybe a minute. Far less painful than trying to sort thru the arc fragments made by dividing.

 

The way I see it, you don't need to sort through those arc fragments. You just need to delete them en masse so that you're left with the three shapes on the right.

[Alfred]

Instead of creating two curves (compound) objects for the larger two, it just duplicates the objects needed to create those curves (compound) objects.

 

Yes, that's why I said

 

I think it should combine objects to leave you with three objects instead of five

[gdenby]

 

 

gdenby, when I place the circle at the bottom of the layer stack, I get just 7 objects after the Divide action. Did you try that?

 

That 1 I didn't try. 

[gdenby]

Yup, Circle on the bottom, 7 curves.

[R C-R]

Alfred, I was just trying to graphically demonstrate that what we think should happen is not what actually does happen.  :)

 

I'm still not convinced that the behavior should be the same for unfilled shapes as for filled ones, or if the expected behavior should create curves objects or compound ones, but I will leave that for the developers to sort out.

[Alfred]

I'm still not convinced that the behavior should be the same for unfilled shapes as for filled ones

 

I'm still not convinced that unfilled shapes should be treated any differently than filled ones. :)