Mastering Layers Q&A

[dgs]

I think what you have here is the difference between a feature and a feature entity. All features have entities, edges, axis, surfaces, points, vertices, etc. A datum plane feature contains a single entity, the datum plane. By searching FOR Datum Planes BY Datum Planes you gathered a set of Datum Plane entities. normally, the plane entities aren't selectable and they aren't able to be hidden on their own. You hide the datum plane feature, not the entity. The 'x\a' icon is the giveaway, that indicates an entity rather than a feature.


By changing it to LOOK FOR Feature, you get the the datum plane feature which can be hidden.


In the context of a Datum Plane feature, the feature/entity distinction is pretty meaningless. However, in the context of an extruded sketchthat contains several circles and thus creates several axis, it's very meaningful. You can grab the axis themselves or all of them by grabbing the feature

 

[gkbeer]

The difference between choosing features and geometric entities is a small but important one.

Consistency in choosing one over the other, is key to having layers that work as expected.

For a datum plane there are some advantages to always putting the feature on a layer.

One is when the layer containing your datum plane features is hidden, and you select a datum plane in the model tree, the feature will highlight on the screen. If the layer contains the geometric entities instead, that highlighting won't occur. The same is true when choosing a layer in the layer tree. Features will highlight, entities won't.

For anyone trying to learn how layers work, or are trying to debug their layers, start by making sure you are only putting features on layers.

When that level of consistency is in place, then you can begin to see where it is useful to have layers with entities.



An advanced example of choosing to put entities on layers:

Consider threaded holes, they have several types of geometric entities, there is an axis, and axis tag, and a quilt that represents the thread.

If you desire the ability to control the visibility of the axis and the threads independently. You will need separate layers for each. In this case putting the hole feature on either of these layers will prevent you from accomplishing that goal.

Look for: axis, look by: feature, type=hole will gather all the hole axis entities.

Unfortunately in WF3 there isn't any way to generate a layer rule that will select the relevant quilts. That would be a case of look for: qulit, Look by: feature.type hole.
At present each quilt has to be selected manually.

 

[Tunalover]

Folks-
None of us want to make a career of managing layers and entities with layers!
All I want to do is delete the layers from a legacy part of assy, import our standard layers, then assign rules to the layers. How is this done simply?
TL

 

[gkbeer]

Simply,

1. Create a rep showing only the subassemblies and parts to be updated.
2. Erase the not displayed items.
3. Delete all the layers
4. Create rule driven layers in the top level assembly. If those rules already exist in the assembly skip this step.
5. Highlight/Select the assembly layers and choose from the "Layer" pulldown "Extend Rules"

 

[Tunalover]

You say "Simply..."

Think a minute about how long this process will take with one part let along an assy with many parts in it. Most companieshave ten or more standard layers each with AT LEAST one rule in it. Some layers may have three or four rules. For example, a layer called GDT might contain feature control frames and geometric datums. If we have a second layer called DEF_DTMS and a third layer called NON-DEF_DTMS then more rules must be written to insure that GDT doesn't contain default datums and created datums. Likewise, DEF_DTMS will need rules to make sure it doesn't contain geometric datums and created datums. And so forth and so on...

 

[verge]

Sounds like you may want to consider simplifying your layer structure. Personally, I think these default Pro-E layers like DEF_DTMS are worthless; I've never been able to keep track of which layers are default or created while I'm within a massive assembly. Nor do I care.


I was actually the one who started Glenn down this perilous path by asking him to explain this technique, and I did so because the default layers in Pro-E aren't useful to me. I am now able to apply this technique to an assembly consisting of thousands of parts in a matter of minutes. Most of the time is spent waiting for Pro-E to delete the old layers.


I still use this technique, even though I now work for a company where no-one else uses it. My models are all in development, so I can change my layers anyway I want without messing anyone else up. The first thing I do when someone sends me an assembly is delete all the layers and create the ones I want. I only want five or six layers for my own use, so this takes me literally less than five minutes, including the time it takes to delete the old layers.


Simplify your own tasks as much as you're allowed to. This technique has saved me many, many hours of work since I began using it. It's worth a few lunch hours' time to figure it out, and it would probably be worth it to try to implement it where you work. I'm working on that here now, and finding that, once people see what it can do, they're pretty easy to convince.


It does make me cringe watching our designers work without ever using layers. They just turn on all of the datums, then use query select to find the one they want. Half the time they're just grabbing one that's in the right place; it's not necessarily the correct datum, just one that happens to be in the right place. This leads to some interesting problems with parts failing. This technique can eliminate this for good.

 

[gkbeer]

You say "Simply..."

Think a minute about how long this process will take with one part let along an assy with many parts in it. Most companieshave ten or more standard layers each with AT LEAST one rule in it. Some layers may have three or four rules. For example, a layer called GDT might contain feature control frames and geometric datums. If we have a second layer called DEF_DTMS and a third layer called NON-DEF_DTMS then more rules must be written to insure that GDT doesn't contain default datums and created datums. Likewise, DEF_DTMS will need rules to make sure it doesn't contain geometric datums and created datums. And so forth and so on...

Yes, Simply


One of the techinques buried in the bowels of this topic suggests creating an empty "def layer" assembly containing all the layers desired, with all the rules desired.
When faced with a problem assembly, open it and wipe the layers, open the "def layer" assembly, add the problem assembly, extend the layers, save the formerly problematic assembly. and erase the modified "def lay" assembly.

While that works, a good set of mapkeys is better.

 

[jelston]

Was it here (this topic)that i saw an answer to getting rid of piping centerlines in drawings?

 

[gristle]

Hi layer gurus.
I have a question for you regarding rules.
Is there a way to select a series of features between two other features, but not by using the feature number collector under history. That works fine as long as you are not adding or deleting features... I thought it may be possible to convert a feature's ID into its current feature number, but I have had no luck.

Thanks in advance.

 

[ddrichards]

I'm happy to see layers being discussed, most of the Pro-E users I been exposed to are of the belief layers are obsolete because you can turn the display off using tool-bar buttons.

 

[jelston]

Was it here (this topic)that i saw an answer to getting rid of piping centerlines in drawings?

Figured it out.

 

[dgs]

Figured it out.

Can you share with the rest of the class?

 

[jelston]

oh, sure. hide the pipe feature at the part level and the centerlines will go away on the drawing (might have to update drawing). To be more specific, you can use a search function like this:


Then hide the layer.


I'm pretty sure someone hasalready posted the answer, but here it is again.
Edited by: jelston

 

[gkbeer]

Hi layer gurus.
I have a question for you regarding rules.
Is there a way to select a series of features between two other features, but not by using the feature number collector under history. That works fine as long as you are not adding or deleting features... I thought it may be possible to convert a feature's ID into its current feature number, but I have had no luck.

Thanks in advance.

I'm not sure what is being asked. Is it seems that you wish to setup a query where one rule selects a specific feature, perhaps by name, a second rule to select another and everything in between. All this while not using fid's or feature numbers in the rule(s).

If you could, then those two features could be moved around in the model tree, causing the contents of the layer to change.

Edit: Didn't think of a way to do that. Would using a test of parent child relations work for your problem?



Edited by: gkbeer

 

[gristle]

Along those lines. I use evaluate features just as a name placeholder in the tree, for example, top housing, bottom housing etc. The names appear at the top of the construction for that specific area. Ideally I would like to specify the feature at the top and bottom of a range to place features between on a layer. I can do it manually with the feature number collector, but have to manually update the feature number range if I add or subtract features that fall within the range.

 

[gkbeer]

I have the solution!

Sounds like you are trying to put a group of features on a layer. Pro/e does that.

Select a range of features then RMB>GROUP. Add the group itself onto a layer. You can revise the extent of the group by dragging items into or out of it. (modeltree) Adding or removing an item from the group will add cause that item to be visible or invisible same as if it was on the layer.

Lots of advantages.
<ul>[*]Putting a group on a layer is as good as putting all the individual features on the layer.
[*]You can compress the model tree by closing up the group. [*]The group can be named usefully. [*]It can be selected and then RMB>Edit will cause all the dimensions pop up to be revised, as opposed to editing dimensions, one feature at a time. [/list]edit: There are layer rules that can automatically add groups to a layer.

PS: you are going to need to learn to use HIDE/ISOLATE instead of HIDE/UNHIDE.


Edited by: gkbeer

 

[gristle]

Hi there, thanks for that reply. I will look into it but from memory in WF2 if a feature in a group fails, all the features in the group are suppressed? Not sure if this is different in the newer versions (need to use WF2 on this project)

 

[gkbeer]

If a feature in a group fails, you get kicked into repair mode and fix it.

It is true that if you choose to suppress the broken feature, the entire group will be suppressed. Resuming the group will toss you into the repair mode again. So fix the feature.

Doesn't seem to be a reason to dismiss using groups, especially if they exactly solve the original problem of putting a sequential set of features onto a layer.

 

[gristle]

You're right there. Just a shame that there is not another way to do this, if for whatever reason groups are not suitable.

 

[handyman2000]

Is there an easier way to make corrections to a mapkey than re-creating every step again? I am used to working with Excel macros and most times can just edit the existing code. The code for the mapkeys found in the config.pro file doesn't seem as straight forward to edit. Is there a different way to edit a mapkey and step through the existing steps to make changes? I am just having trouble creating an effective mapkey to reset all the layers and cleanup parts so that in big assemblies it is not a complete mess of datum features.