Jump to content

 
Photo

dissolve issue

- - - - -

  • Please log in to reply
11 replies to this topic

#1
Leah Brooks

Leah Brooks

    Newbie

  • Validated Member
  • Pip
  • 4 posts
  • Canada

Hello --

I'm using ArcGIS 9.3, and I am having difficulties with the dissolve tool. I have a map with many voting precincts (approx 900) that I want to dissolve into council districts (8 of them).

I use the dissolve tool and choose the council district as the dissolve field and get 8 polygons. However, the borders of the new polygons don't follow the borders of the maximal set of original polygons in each council district. In particular, the borders are wiggly where the original borders are more or less straight.

I am attaching a pdf that shows the original map, the dissolved map, and a zoomed-in portion of the two overlaid to get a sense of the problem.

Is there any way to fix this problem?

Thanks very much,
Leah Brooks

Economics
University of Toronto

Attached Files



#2
Hans van der Maarel

Hans van der Maarel

    CartoTalk Editor-in-Chief

  • Admin
  • PipPipPipPipPipPipPip
  • 3,868 posts
  • Gender:Male
  • Location:The Netherlands
  • Interests:Cartography, GIS, history, popular science, music.
  • Netherlands

Hi Leah,

Could you perhaps send us the original data as well (shapefiles) so that we can have a go ourselves?
Hans van der Maarel - Cartotalk Editor
Red Geographics
Email: hans@redgeographics.com / Twitter: @redgeographics

#3
Leah Brooks

Leah Brooks

    Newbie

  • Validated Member
  • Pip
  • 4 posts
  • Canada

Hi Leah,

Could you perhaps send us the original data as well (shapefiles) so that we can have a go ourselves?


I'm very happy to! They are attached.

The initial file is
p99_...
and I dissolved by council_di

The dissolved file with the strangely wiggly boundaries is
la_p99_council_...

Attached Files



#4
Matthew Hampton

Matthew Hampton

    Hall of Fame

  • Moderator
  • PipPipPipPipPipPipPip
  • 1,303 posts
  • Gender:Male
  • Location:Portland, Oregon
  • Interests:Playing in the mountains and rivers.
  • United States

Wow. That's messed up!

I did a quick check and found the following info from ESRI.

It looks like you might need to do them in small batches?

Ironically I performed the Dissolve flawlessly in ArcView 3.2 ;)

Do you have access to that?

co-cartographic creator of boringmaps.com


#5
Leah Brooks

Leah Brooks

    Newbie

  • Validated Member
  • Pip
  • 4 posts
  • Canada

Wow. That's messed up!

I did a quick check and found the following info from ESRI.

It looks like you might need to do them in small batches?

Ironically I performed the Dissolve flawlessly in ArcView 3.2 ;)

Do you have access to that?


Thanks for the reply. The librarian at the U of T had success using merge for each council district individually, but I have to do four more of these maps (with 8 districts per map), and that seemed a bit nuts!

In the ESRI help you link to, I notice that their dissolved polygons follow the original (non-wiggly) borders!

I do have an old ArcView, so I will try that.

But in principle, this dissolve problem just seemed very wrong, and I was hoping that someone might know what was going on.

Thanks much,
Leah

#6
Hans van der Maarel

Hans van der Maarel

    CartoTalk Editor-in-Chief

  • Admin
  • PipPipPipPipPipPipPip
  • 3,868 posts
  • Gender:Male
  • Location:The Netherlands
  • Interests:Cartography, GIS, history, popular science, music.
  • Netherlands

It almost looked like there was a dissolve and some generalisation going on. Really weird.

I threw the data into FME and came up with this
Attached File  output.zip   175.09KB   21 downloads

9 areas instead of 8, FME's Dissolver won't automatically aggregate, and the resulting boundaries follow the original ones. So it's not the input data... Oh, and as an added bonus, the resulting output attribute has the same name as the original input one :rolleyes:

I'd be happy to do the other maps as well if you want.
Hans van der Maarel - Cartotalk Editor
Red Geographics
Email: hans@redgeographics.com / Twitter: @redgeographics

#7
Matthew Hampton

Matthew Hampton

    Hall of Fame

  • Moderator
  • PipPipPipPipPipPipPip
  • 1,303 posts
  • Gender:Male
  • Location:Portland, Oregon
  • Interests:Playing in the mountains and rivers.
  • United States

In the ESRI help you link to, I notice that their dissolved polygons follow the original (non-wiggly) borders!

I do have an old ArcView, so I will try that.

But in principle, this dissolve problem just seemed very wrong, and I was hoping that someone might know what was going on.

Thanks much,
Leah


The text in the ESRI help that tipped me off to potential difficulties is as follows: "Dissolve can create very large multipart polygon features. This is especially true when using a small number of Dissolve Field(s) or when dissolving all features into a single feature. Very large features may cause display problems and/or have poor performance when drawn on a map or when edited...Dissolve limits the creation of extremely large features based on the relation of feature size to available memory. In such cases, smaller more manageable features will be created. Warning messages will be used to indicate this occurrence."

In my mind this could be boiled down to "Things get funky when you work with larger multi-part features - proceed with caution."

The 9.3 help for Dissolve (I linked you to 9.2 help) includes mention of an adaptive tiling algorithm used for large data sets. I wonder if this is causing problems?

Good luck!

co-cartographic creator of boringmaps.com


#8
sitesatlas

sitesatlas

    Master Contributor

  • Validated Member
  • PipPipPipPip
  • 182 posts
  • Gender:Male
  • Location:Madrid, Spain
  • Spain

Hi Leah,

I was able to dissolve just fine in Manifold. There is a precision setting that is normally set to 0.000001 degree in Manifold, but if it is made larger, the data will be simplified when dissolve or other operations are performed. This sometimes creates irregular shapes like the ones you got.

Could there be something similar in ArcGIS?
Michael Borop
World Sites Atlas
http://www.sitesatlas.com

#9
Hans van der Maarel

Hans van der Maarel

    CartoTalk Editor-in-Chief

  • Admin
  • PipPipPipPipPipPipPip
  • 3,868 posts
  • Gender:Male
  • Location:The Netherlands
  • Interests:Cartography, GIS, history, popular science, music.
  • Netherlands

The text in the ESRI help that tipped me off to potential difficulties is as follows: "Dissolve can create very large multipart polygon features. This is especially true when using a small number of Dissolve Field(s) or when dissolving all features into a single feature. Very large features may cause display problems and/or have poor performance when drawn on a map or when edited...Dissolve limits the creation of extremely large features based on the relation of feature size to available memory. In such cases, smaller more manageable features will be created. Warning messages will be used to indicate this occurrence."


IMHO it shouldn't. If I'd want large multipart features, I'd aggregate... I do realise there is a difference in the terminology used between Arc and FME (and Manifold).
Hans van der Maarel - Cartotalk Editor
Red Geographics
Email: hans@redgeographics.com / Twitter: @redgeographics

#10
Crischan

Crischan

    Master Contributor

  • Validated Member
  • PipPipPipPip
  • 199 posts
  • Gender:Male
  • Location:Eberswalde, Germany
  • Interests:Spatial Information Design
  • Germany

Hi,

I have a ArcGIS ArcInfo 9.2 workstation and gave it a shot: I noticed that the shapefile has no CRS information but seems to be GCS_WGS_1984...

First try: Same wiggly results as you.
Assigning the data frame a GCS_WGS_1984 CRS - same result.
Assigning the shapefile a GCS_WGS_1984 CRS - Yeah, it works!

We have made that experience before - if on-the-fly projection or no projection is involved many tools screw up badly, mostly without giving a warning. We also noticed that the erroneous behavior changes from service pack to service pack and, of course, from 9.2 to 9.3... :)

Thank god it's such a bargain of a software... ;)
Crischan Wygoda
http://wygoda.net

#11
James Hines

James Hines

    James Anthony Hines

  • Validated Member
  • PipPipPipPipPipPip
  • 537 posts
  • Gender:Male
  • Location:Centreville, Nova Scotia
  • Interests:Cartography, Philosophy, Politics, Psychology, Economics, Occultism, Spiritualism
  • Canada

I know the problem has been fixed but I decided to take a crack at it in Manifold & I had no problem with the dissolve. However I will say I tried one time with population values in an entirely different shape file & had the same problem. Oh well it wasn't for a project.

"There is much beauty that we fail to see through our own eyes teeming with life forms that give us that perception of our reality.  Leaves on the trees blowing gently in the wind, or scarily, the waves pounding through high surf, or lightly on a warm summer’s day; that opportunity to sit or swim in the water on a white beach.   That comfort to shout, “The universal conscious do you hear me?  I am alive, guide me dear logos towards the path of rightnesses.”  Earned what has been kept, no longer to be absorbed into a life filled with cold damn winds and  that stubborn fog clouding  my vision with nothing but darkness."


#12
Leah Brooks

Leah Brooks

    Newbie

  • Validated Member
  • Pip
  • 4 posts
  • Canada

Hi,

I have a ArcGIS ArcInfo 9.2 workstation and gave it a shot: I noticed that the shapefile has no CRS information but seems to be GCS_WGS_1984...

First try: Same wiggly results as you.
Assigning the data frame a GCS_WGS_1984 CRS - same result.
Assigning the shapefile a GCS_WGS_1984 CRS - Yeah, it works!

We have made that experience before - if on-the-fly projection or no projection is involved many tools screw up badly, mostly without giving a warning. We also noticed that the erroneous behavior changes from service pack to service pack and, of course, from 9.2 to 9.3... :)

Thank god it's such a bargain of a software... ;)


Thank you all so much! I agree that when I assigned a projection to the shapefile, the dissolve worked appropriately. I will try to keep that in mind for future issues.

And thanks for all the other replies -- pointing out about how the algorithm worked -- I naively assumed that it did something much simpler!, and offering to help in Manifold.

All the best,
Leah




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users

-->