Skip to main content
  • EN

Hello!

I have SHP file with the bunch of streets. There are many streets with the same name which is ok - streets with the same name can exist. But also there are some streets with the same name which are actually one street but entered as two different features instead of the one feature. So those streets have two geometries and the same name and are actually one street. I want to isolate (select) such streets. Is there any suggestion how can it be done in FME?

So, for example, street with the name "Oak street" can have splitted geometry in one location but also there can be many other "Oak streets" in other locations (some of them regular with only one geometry, but also maybe other with also splitted geometries).

So, the idea is to select streets with the same name which are in the neighborhood of 100m (for example 100m - the assumption is that those are most likely streets with the splitted geometry).

I would appreciate any help and thanks in advanced!

Here's one suggestion. Use a Bufferer to set a 100 m buffer around streets, grouped on street names. The buffers get the street name. <redundant>Then use a Dissolver to make one big buffer around each street, by grouping on street name. (Send the streets both through the Bufferer and past it, to keep both.)</redundant> Then put a Counter on the buffers to get a unique combination of street name and buffer count. Finally use a LineOnAreaOverlayer to match the streets to their buffer and make them inherit the buffer's count attribute. Now you should have street features with a unique combination of street name and count number. This should work for a scenario where you don't want to aggregate or join the lines, but just want to give them a unique, searchable attribute.

(Edit: Marked the redundant part of my suggestion, but kept it in there to make the following answers and comments understandable.)

Thank you very much for your suggestion! I really appreciate you help.

I did almost everything as you suggested and I'm now analyzing the results (trying to confirm that the suggested logic can be applied to my data).

One thing I didn't understand is this step: "Then use a Dissolver to make one big buffer around each street, by grouping on street name. (Send the streets both through the Bufferer and past it, to keep both.)"

I didn't understand why I need dissolver because bufferer already dissolves buffers by street names - If I understood the help manual. And the instruction in parentheses I didn't understand - it seems unnecessary but I would like to discuss because maybe I didn't fully understand something.

So, I did it like this:

  1. I used bufferer to create 100m buffer polygon around all streets with Group by parameter on street name >>> as a result, all buffers that overlaps and have the same street name are dissolved into one buffer (am I right?)
  2. I added a counter to produce unique identifier for the each buffer
  3. I used the LineOnAreaOverlayer to intersect streets and buffers so that each street gets the buffer ID
  4. I selected streets which belong to the same buffer, i.e. have the same buffer ID and have the same name.

I can not find a flaw in this 4 steps, but maybe I'm missing something and should also include those extra steps you suggested (dissolver and parentheses...)

 

 

 

Thank you very much for your suggestion! I really appreciate you help.

I did almost everything as you suggested and I'm now analyzing the results (trying to confirm that the suggested logic can be applied to my data).

One thing I didn't understand is this step: "Then use a Dissolver to make one big buffer around each street, by grouping on street name. (Send the streets both through the Bufferer and past it, to keep both.)"

I didn't understand why I need dissolver because bufferer already dissolves buffers by street names - If I understood the help manual. And the instruction in parentheses I didn't understand - it seems unnecessary but I would like to discuss because maybe I didn't fully understand something.

So, I did it like this:

  1. I used bufferer to create 100m buffer polygon around all streets with Group by parameter on street name >>> as a result, all buffers that overlaps and have the same street name are dissolved into one buffer (am I right?)
  2. I added a counter to produce unique identifier for the each buffer
  3. I used the LineOnAreaOverlayer to intersect streets and buffers so that each street gets the buffer ID
  4. I selected streets which belong to the same buffer, i.e. have the same buffer ID and have the same name.

I can not find a flaw in this 4 steps, but maybe I'm missing something and should also include those extra steps you suggested (dissolver and parentheses...)

 

 

 

You're right. My suggestions were from the top of my head, so a slight redundancy crept in there. 🙂 Your four steps should do the trick. Have you reached your wanted result?

You're right. My suggestions were from the top of my head, so a slight redundancy crept in there. 🙂 Your four steps should do the trick. Have you reached your wanted result?

Yes, I believe so. With my test data with only 10 streets it works great but I have more than 100000 streets so I wasn't sure if I have missed something important in your suggestions which doesn't affect the results of my test data but could give false results in my real data (some special cases which haven't come to my mind while preparing my test data).

So, once again, thank you very much!

M

Reply


Community Stats

31,821
Posts
120,861
Replies
39,452
Members

Latest FME

Terms & ConditionsCookie settings
A product of
Safe Software respectfully acknowledges that we live, learn and work on the traditional and unceded territories of the Kwantlen, Katzie, and Semiahmoo First Nations.