I'm not really sure how I ended up with Bezier curves in our parcel fabric, however, I recently came across an area that contained quite a few of them. They seemed to cause a ton of weird issues while editing. I manually converted them all to circular arcs, and that resolved the problem, but now I'm wondering if there are any more in my data. Is there an easy way to find and remove them? I would have thought that running the "Simplify by Straight Lines and Circular Arcs" tool did that, but it seems not because I ran my entire dataset through that prior to migrating to the parcel fabric.
The beziers may have been introduced as a result of running the GP Smooth Line tool, or by using the editing tool called Generalize with the Smooth option. The SLACA tool intentionally does not alter the geometry of features with parametric segments such as beziers, circular arcs and elliptical arcs, as this would be considered to be data corruption by customers who'd like to keep their beziers intact.
To answer your question, I have not found an out-of-the-box way to detect bezier curve geometries in Pro. I was looking into the GP tool called Calculate Geometry Attributes, but it does not report on specific parametric geometry types. That may be a good Idea to add as an enhancement for that GP tool.
There is an add-in (currently unpublished) that I can share with you. It will select features (within the active map's extent) that have geometry segments that are beziers. I'll direct message you to find out more about your ArcGIS Pro release etcetera.
Matthew and Tim:
Are bezier curves supported in the Parcel Fabric ? Is there a difference between a true curve and a bezier curve?
I ask because another thread indicates major issues a user is having with bezier curves in the Parcel Fabric.
That is actually my idea thread as well! Bezier curves are defined by a set of control points. Here is an example of lines with Beziers. True curves are curves that are defined by coordinate geometry like direction, radius, and arc length.
In my case, Bezier curves completely broke the fabric when working on parcels that contained them. Thankfully, with Tim's help, I was able to clean my dataset on them. Based on my conversations with Tech support, I believe ESRI is updating their documentation and looking into a way to implement a native function for tracking down these curves since they seem to cause issues with the fabric.
Michael and I work together, and he's the one who alerted me to your other post. Our problem is two-fold, with one issue being the ready identification of any Beziers (we're thinking we've got a mixed bag), and the other being a general reluctance (to put it mildly) to simply generalize all such data, as that defeats the purpose of ever having improved the data w/ Beziers in the first place.
Also, of the (4) tools mentioned by Ayan Palit in the other post, all of them exclusively seem to convert curvilinear data to splined, straight segments, not true curves. In our aforementioned prior cleanup from several years ago, we opted to manually replace the ugly, splined, straight-segmented "curves" by hand, because the output of the Simplify by Straight Lines and Circular Arcs tool was awful - resulting in kinked, non-smooth strings of segments. I just hate the idea of regressing our data because of the shortcomings of the Fabric model itself.
I know what you mean. After tweaking the settings of the SSLCA tool many times, I was able to get it to do a pretty good job for the most part, but there are a few areas where I definitely noticed some of the kinking and weird curves that you mentioned. It's not a lot, but definitely enough to be an annoyance.
The short answers to both questions is yes. Beziers are different to circular arcs, and they are a supported geometry type for features in a parcel fabric. There are, however, some limits to using beziers in the current release, described below.
Feature geometry is built up from segments. A segment can be a straight line, or a parametric geometry. A parametric geometry can be a circular arc, an elliptic arc, or a bezier. These parametric geometries are referred to generically as "curves", but when used in the context of parcel data and COGO we are, 99.9% of the time, referring to circular arcs. In the COGO world "true curves" and circular arcs have become synonymous.
Limits of beziers when using Build command:
Beziers may be used in the fabric model to represent natural boundary features, however we have seen cases where straight lines have been turned into bezier curves even though they are still geometrically straight. Beziers should be reserved for natural boundaries. If your fabric has very large natural boundary features with many bezier curve segments there are known limits that cause performance problems when running the Build command, for example. These issues are being addressed and fixed for the next release of Pro. (3.1) We are also investigating the feasibility of providing a patch for this on earlier releases.
What about transition curves and spirals?
The other curve we see in COGO workflows is the "transition curve" used for highway and rail centerlines. This incorporates the use of spirals, and combining them with circular arcs and/or straight lines to form smooth, tangential transitions to represent long linear features. Spirals cannot be represented parametrically in Pro; they are handled differently and are represented as a densified polylines. (For more specific info see this topic in the Concepts documentation for the COGO SDK)
I'd recommend contacting Esri tech support and reference this post and the case number as per our e-mail correspondence. That would help add weight for us to get a patches based on the release you're using.