Nonassociative hatches

Sascha Boskovski

Hello.

I understand, that I can´t stretch a nonassociative hatch via picking and moving a corner of the surrounding polyline as the two are not connected.
But why can´t I stretch a nonassociative hatch via command "stretch"?

Have a look at the attachment.
If the locker has to be deeper, and the hatches are nonassociative, it´s a real mess to stretch as the hatches stay like they are.

Thank you.

Sascha

Joe Dunfee

 I have also often been disappointed when a hatch that started out as being associative, ends up loosing that association.  I haven't studied it to try to figure out what causes it to loose associativity (is that a legitimate word? - well it is now.).  My first guess is that a set of lines and associated hatches will not retain their associativity when they are copied.

Has anyone here studied the behavior and can state what breaks associativity?

-Joe

Roy Klein Gebbinck

@Sascha:
_HatchGenerateBoundary
Or:
_-HatchEdit > _Associate

@Joe:
Of the top of my head I know of one issue in V14, the version you are using, but it has been fixed in later versions.
Here is a small test to demonstrate it:
1. Draw a rectangle (polyline).
2. Hatch the rectangle using the 'Pick points' method.
3. Move only the rectangle so that the point you picked in step #2 no longer falls inside it.
4. Result in V14: The hatch becomes disassociated and is left behind.

General advice:
When hatching a closed boundary (polyline, circle etc.) you should always use the 'Select boundary entities' options. It is faster and more accurate.
When using the 'Pick points' method keep in mind that the points you pick are stored and will be reused when the hatch is regenerated. Choosing these points wisely can sometimes be important.

Roy Klein Gebbinck

... Oops that should of course be: "Off the top of my head ..."

Anthony Apostolaros

Sascha:  Why is it necessary for the hatches to be non-associative? If they were associative, that would eliminate the problem you described. You would be able to stretch the depth of the locker array including the hatches. Would hatch associativity cause another problem? If so, perhaps something could be done about that other problem.

Sascha Boskovski

Hello Anthony.

Lets look at the situation with the lockers.
I create e.g. a sidewall and put in a hatch. The hatch is associative.
If I have to cut out a part of this element by triming the polyline, and rejoining the new lines, the hatch looses the associativity.
From this point on I have a lot of work to stretch things.
See attachment.

I just found out, that the hatch gets nonassociative after joyning the polylines.
Is that a bug?

But anyways: where is the sense that I can´t stretch a nonassociative hatch?

Thank you.

Sascha

Screenshot_9.jpg

Roy Klein Gebbinck

@Sascha:
I have tested this scenario on V15 (AFAIK this is the version you are using) and V16 and I don't have this issue.
Three things are important:
1. I have used the 'Select boundary entities' option when creating the initial hatch (see my previous post).
2. I have not trimmed the hatch. This would make no sense here.
3. When joining the two polylines with _Pedit I of course choose the entity that was hatched initially as the polyline to join segments to.
4. Result: The hatch is and stays associative.

If you have used the 'Pick points' method the hatch loses its associativity when a gap is introduced in the boundary. Which is perfectly logical.

Anthony Apostolaros

But anyways: where is the sense that I can´t stretch a nonassociative hatch?

I think it's that you can't stretch a hatch of any kind, just as you can't stretch a circle or ellipse or block insertion. I think the associative hatch is simply adjusting when its associated boundary is stretched, so it appears to be altered by the Stretch command. If I draw a rectangle with an associated hatch, and then draw two lines inside the rectangle, and then execute the Stretch command with a Crossing box enclosing the first line but not enclosing any side or vertex of the rectangle, and crossing the second line and the hatch, the first line moves and the second line stretches, but the associative hatch doesn't change.

crossing.png