Union tool running for over 24 hours

2579
8
Jump to solution
12-19-2017 07:05 AM
MorganHarris
New Contributor III

I'm running the union tool on a large stream layer (all of western NC) to merge waterbodys (lakes and ponds) with my stream network (both features are polygons). I tried the merge tool first and it ran for over 24 hours, being stuck at 100% complete for 16 of those hours, so I cancelled it. Now I am trying the union tool since it subdivides features for processing. Its been running for about 24 hours now and it has been saying "processing tiles" in the status and keeps going up to varying percentages (usually less than 50) and then jumping back down to 0 without ever reaching 100% on any one process. 

Is this the normal way that Union runs through the tiles, or is something going wrong here that warrants cancelling this operation? I'm wondering because I don't want to waste days of processing on something that isn't going to work.

0 Kudos
1 Solution

Accepted Solutions
XanderBakker
Esri Esteemed Contributor

You are right that dissolving the features after (or during the buffer) would probably better. This will be time consuming too. 

The merge should be faster when you do not use the complex multi part feature. What we don't know is if the current process is almost finished. Also canceling an operation could be time consuming and sometimes might require killing the process. 

View solution in original post

0 Kudos
8 Replies
XanderBakker
Esri Esteemed Contributor

Can you indicate how many features you have in both featureclasses? 

0 Kudos
MorganHarris
New Contributor III

The number of features isn't that huge. I've actually reduced the stream layer to only 2 features. The waterbody layer has 415; however, I'm sure the number of vertices is pretty huge. I want to look at how many vertices there are, but I'm concerned that would interrupt the tool.

0 Kudos
XanderBakker
Esri Esteemed Contributor

I'm sure that merging 2 and 415 features should not take that long (should be seconds or minutes at most). In what format do you have the information?

0 Kudos
MorganHarris
New Contributor III

They are both file geodatabase feature classes. More about the size though, before I reduced the stream layer to only 2 features it was 163500, just to give you an idea of the potential number of vertices, if that matters. 

0 Kudos
XanderBakker
Esri Esteemed Contributor

So it is a multi part feature. Then it does make sense that it takes long. It is not recommended to create multi part features with that amount of vertices (and parts). Do you really need to handle it a a single feature?

0 Kudos
MorganHarris
New Contributor III

Down the road I think it will matter because I'm planning to buffer these and I want the buffer to be a single feature, but I suppose I could get that result just by making sure the buffer is dissolved into a single feature instead? 

So it sounds like I'm doing things in the wrong order with this layer. Do you think if I don't make it a multipart feature that it would run smoother?

0 Kudos
XanderBakker
Esri Esteemed Contributor

You are right that dissolving the features after (or during the buffer) would probably better. This will be time consuming too. 

The merge should be faster when you do not use the complex multi part feature. What we don't know is if the current process is almost finished. Also canceling an operation could be time consuming and sometimes might require killing the process. 

0 Kudos
DanPatterson_Retired
MVP Emeritus

It is the number of points and edges that need to be dealt with.  Complex shapes with many vertices (hence edges) will increase processing time.  Sometimes tiling  using Dice or Split since the sum of the parts may be less than the whole due to memory issues and read/write problems/slowdowns when working with large datasets