Select to view content in your preferred language

Joining points to segments more computationaly effieciently?

1596
5
08-15-2016 04:14 AM
BachirHamadeh
Deactivated User

I have over two million crashes located as points through GPS coordinates in the state of Pennsylvania. I also have over a hundread thousand road segments in the state and would like to know the number of crashes per segment. I am right clicking the segment feature class and using join with the point feature class and choisibg summarize sum to do this. I did the analysis on a single county and it worked except when i extrapolated the time to do this for the entire state, i came out with several several weeks with non-stop processing. Is there any way I can process this faster without reducing the number of segments i am merging the crashes to? 

0 Kudos
5 Replies
NeilAyres
MVP Alum

See my comment on your other thread.

0 Kudos
DanPatterson_Retired
MVP Emeritus

Yes

Neil... where is the other thread, ... don't want to duplicate answers

0 Kudos
NeilAyres
MVP Alum

Dan,

sorry but the author has posted this twice, the other one is in Managing Data.

Don't know how to cross link the 2.

0 Kudos
RebeccaStrauch__GISP
MVP Emeritus

Neil Ayres‌  Easiest way to mention a cross post is just to mention it is a duplicate hse and include the link to the other (or whichever already has  comment(s)/answer)  e.g.

https://community.esri.com/thread/181427-spatial-join-more-time-effeciently 

To post and include two locations, it's a s easy as use the at and space/place name, eg

https://community.esri.com/community/gis/managing-data?sr=search&searchId=30e83fec-1fb8-46e6-aa64-09...

The original posted has the option of deleting this thread and adding this space  https://community.esri.com/community/gis/analysis/spatial-statistics?sr=search&searchId=8e90ea5c-1a4... to the other thread...to keep it in one

Just an fyi for future.

0 Kudos
DanPatterson_Retired
MVP Emeritus

Your point pattern is important, if it is a spatial join (ie you have the coordinates of the incidents) then you need to perform some sort of spatial aggregation first.  If the coordinates were duplicates, then you could summarize that way.  If the locations are dispersed along a road stretch, then you could use an aggregate points using either the polygon or its centroid prior to a spatial join.

There are many other options, but your task is to aggregate either/or/both sets of data.  Comparing every point to every segment in the standard interface (even with its underlying optimizations) is a waste of time and won't reveal any more information in most instances.

0 Kudos