Problems with Intersect tool in 10.1??

3582
6
11-01-2012 10:23 AM
by Anonymous User
Not applicable
Hi all,

I have a script that I run every night that recalculates CSR values for ag land parcels since there are boundaries that change almost every day.  I have not had any problems until I upgraded to 10.1 two days ago.  My script is failing at the Intersect tool every time and I am getting a strange error:

Error Info:
ERROR 999999: Error executing function.
The table was not found.
The table was not found. [Soils_by_parcel]
The table was not found.
The table was not found. [Soils_by_parcel]
Invalid Topology [Topoengine error.]
Failed to execute (Intersect).

I never got this message before and the script ran fine without any problems with v10.  Is this a bug with 10.1? Is anyone else having this problem?  I have read that the intersect tool may crash on large datasets (which the the soil fc is huge), but it worked fine when I had Arc 10 (has been running every week night for the last 3 months with no problems!).  I have changed the XY tolerance, disabled background processing, repairing geometry, downloaded service pack 1, tried it by script, tried it in ArcMap and it STILL fails.

Both Feature Classes are in a file geodatabse and I am running Arc 10.1 (SP1) on windows XP.
0 Kudos
6 Replies
KenHartling
Esri Contributor
At 10.1 we've found that a number of changes to how memory is being used by subsystems we rely on can affect some mid to large size overlay processes in low memory situations (XP is fairly limited in the amount of memory it can use).  We are working on a number of workarounds for 10.1 SP2 to try and help in these situations.

One possible workaround is to try your Intersect operations on a Windows 64bit machine.  If it does not work in the foreground there, try installing 64bit Background and running it using BG64 on the Windows 64bit machine.

If you do not have a machine to try this on, we can try it here if you provide your data and exact parameters used to run Intersect. I can send you instructions on how to upload your data to our secure FTP site if you'd like us to try it here.

Thanks,
Ken
Senior Geoprocessing Product Engineer
0 Kudos
by Anonymous User
Not applicable
Hi Ken,

Thanks for the reply.  I chatted with an Esri Analyst about this and it was suggested that I divide the area (soil boundaries in my case) up into smaller chunks and then try to intersect.  When this did not even work, I discovered that the feature class somehow got corrupted sometime around the time when I upgraded to 10.1.  I then downloaded a new soils file from the new Esri SSURGO soil mapper (very nice I may add!) and used this to intersect with my parcels instead of the soils FC and it worked fine.  So it seems I was just the victim of a corrupted FC and there was not any problems with my upgrade to 10.1.  I just assumed it was because of the upgrade because my script started failing the day I moved to 10.1.

My script is back in business now with no errors.  I also want to add that I am very pleased with the performance in 10.1.  The script used to take an 90 minutes to run but I am now using da.UpdateCursors instead of the field calculator and the script now completes in only 15 minutes! Very cool :cool:.
0 Kudos
CaitlinJensen
New Contributor
Hello,

I am running into the same error as Caleb1987 when attempting to intersect a fishnet with a line dataset. The original line dataset is extremely large, so I have been sizing it down to many subsets to expediate processing.  This worked fine for the first 8 batches, however for some reason it will not work now. I have tried running it through both ArcMap and Catalog (v10.1), with background processing enabled and disabled, and also with the service pack and 64bit background processing. I'm not actually getting that error message anymore, but the process just gets stuck at "Assembling Features" and doesn't progress. I have let it process for over 24 hours, and still I return to the office to see it stuck at the assembling features stage. I've tried running this intersect multiple times, and after cancelling the process and opening the attribute table of the output feature class, it always has exactly 50,000 records. This makes me think it is a systematic problem rather than a problem with the data itself? This subset is no larger than the other subsets that I have successfully processed, so I am confused why it is particularly troublesome.

Any suggestions??

Thank you!
Caitlin

P.S. - I'm using a computer with 12 GB of RAM, Intel(R) Core i7 processor.
0 Kudos
KenHartling
Esri Contributor
Hi Caitlin,
It may be how complex the data is rather than how many features.  Are you using FGDB?

First, have you looked at this blog and gone through the suggestions to make sure everything is setup for you to be successful with large overlay operations?

If so and it's still a problem could you share the data with me?  I'll also need the exact parameters you used in the Intersect tool when you encountered the problem.

Here's how to upload your data securely to our ftp site:

"You can place data on the /pub/incoming location on our FTP site. However, no data can be downloaded from that location except by internal Esri users using a special user name and password.

Login to ftp://ftp.esri.com/ using the anonymous login and providing you e-mail address as the password if prompted (Note: Most browsers and FTP clients will log in using the anonymous account automatically).

Once connected, navigate to the pub/incoming folder, create directories as needed and upload you content."

Thanks,
Ken
0 Kudos
CaitlinJensen
New Contributor
Hi Ken,

Thanks for your quick response. I read through that blog and believe that I have adhered to all those requirements. I'm trying to share the data with you, but there's no pub/incoming folder for me to upload it to...the only folders there are outgoing/, sremote/, and staff/.  Where should I put my data?

Thanks,
Caitlin
0 Kudos
KenHartling
Esri Contributor
Looks like we are phasing out our ftp site in place of an easier to use system.  Please email me so I can get your email address and start the process.
Thanks,
Ken
0 Kudos