Latest Contributions by StevePeaslee

Thanks Jill! Just have to be patient I guess. -Steve ... View more

Hi, I just started seeing similar issues when running an ArcSDE-to-FGDB export script. I hadn't run it for a couple of months, but yesterday I was running a small export of about 60 tables and 5 featureclasses from our ArcSDE 10.1 database (SQL Server 2008R2). Usually the failure was during the Copy_managent command. There was a strange pattern. I first experienced the problem when running it on my laptop. When exporting to a FGDB under a folder on my C drive, it would fail at different times, copying different tables. When I changed my output location to another (slower) internal drive C: it would work every time. The failure was during Copy_management and the error message was not helpful. After each failure I would check the input table and output geodatabase and could find no problems, everything seemed to be fine and I could manually use Copy_management without problem. My laptop drives are C: Micron SSD and 😧 Hitachi TravelStar ATA 500GB Next I moved operations to another higher performance desktop computer. I saw very similar problems, except that the first failure to export to a FGDB under C:\Geodata occurred near the end of the process. All of the tables and featureclasses were successfully copied and the script was in the process of recreating the relationshipclasses. The error message indicated that the output FGDB was read-only. Subsequent attempts to export to c:\Geodata failed after a couple of minutes while copying tables. It did not fail on the same table each time. As on my laptop, I tried changing my output location to another internal E: drive and it worked perfectly every time. Both my C and E drives are Intel 300GB SSDs. About 5 months ago I ran all of these exports hundreds of times for terabytes of data without seeing any issues like this. My current theory is MS Endpoint Protection or a Windows 7 update. We didn't have EndPoint 5 months ago. The C:\Geodata folder is a location that supposed to be excluded from scanning so my results are opposite of what I would have expected, but I'm not sure if the folder exclusions apply to the 'Real time protection' process. We don't have any FGDB file extensions or SQL server database file extensions in our exclusion list. I've been trying to get IT to add those but no luck so far. Next step is ESRI support. It would be helpful if ESRI would put out a white paper or technical bulletin on the subject of virus scan exclusions. An official document would smooth the way tremendously. -Steve ... View more

I'm not sure that I can call this truly solved, but as a workaround, I found that if I converted "E:\Temp" to an ArcInfo workspace with INFO directory, the polygon to raster conversion seems to work fine. -Steve ... View more

Jocelyn, There can be legacy restrictions involved when the output format is GRID. If you are having problems that only occur when the output location is a specific drive or folder, then my first thought would be to check for a path that cannot be a valid ArcINFO workspace. The complete list is documented somewhere (http://webhelp.esri.com/arcgisdesktop/9.2/index.cfm?TopicName=About_the_ESRI_grid_format lists some), but the first things I look for are spaces in the path or strange characters. I also try to keep paths short. Another problem I have run into is when a conversion tool automatically names the output GRID using the same filename as the input. If you have a long shapefie name that can cause the tool to fail with a useless error message. This may only apply to older versions of ArcGIS. Some of these problems have been getting better with ArcGIS 10.1. -Steve ... View more

dmhoneycutt, Thanks. I tried the FIDSet method and got some interesting results using ArcGIS 10.0. When FIDSet was used to get a count for the selected set, it worked really fast when compared to GetCount. For some reason I was unable to get FIDSet to work against a featurelayer containing the entire featureclass (3.3 million records). I was getting back an empty string. Not at at all what I was expecting. There must be something wrong with GetCount in ArcGIS 10.0. I decided to try GetCount one more time, just to get the total number in the featureclass and surprisingly it worked fine. So for now I am using a combination of FIDSet to get the number selected and GetCount for the total number of features in the featureclass. It turns out that my original slow performance was completely due to using GetCount on the featurelayer. I wasn't able try the da cursor since my ArcTool has to be compatible with both 10.0 and 10.1. Thanks anyway for the suggestion. ... View more

Hi Matt, I have used JoinField with Python script tools for 9.3.1, 10.0 and 10.1 SP1. In most cases I used them for featureclass to standalone table 'joins' or raster to standalone table 'joins' within a file geodatabase. I usually don't use JoinField in a table to table join and never across two different workspaces. The only issue I've ever had was at 9.3.1 I had to select all records in the target layer or table to get JoinField to work properly. Otherwise I ended up with the new field(s), but NO data. No error messages were ever generated when it failed either. I haven't had any problems with 10.1, so I dropped the 'select all process' from my scripts. I know there was at least one thread in the forums with folks getting partial data population. I that there was a bug report for this issue, but I never saw that kind of behavior. It was all or nothing for me. Another tidbit, I experienced the same behavior whether I ran JoinField from a script or from ArcToolbox. It is easy enough to test using 'SelectLayerByAttribute'. -Steve ... View more

Hi, In ArcGIS 10.1, I've run into the same problem as kpeter (also SQL Server, Identify column as primary key). Does anyone have a workaround that will allow the table to be exported with a new OBJECTID column instead of changing the type of an existing column from integer to OID? Depending on which geoprocessing method you use to export the table, I've even seen where the original column gets renamed to OBJECTID. That's not cool. -Steve ... View more

Matshopp, I do not believe that any geoprocessing methods are exposed that will allow a user to permanently change a geodatabase field alias. I'm not sure why, it seems like it would be a very useful geoprocessing tool. Hopefully someone else can tell me that I am incorrect. For table aliases, you can use the arcpy.AlterAliasName(table, alias) method. I think this is new at 10.0. In 9.x geodatabases, it was possible (but definitely NOT recommended) to alter aliases by editing the GDB_FieldInfo table. If done incorrectly you could trash the geodatabase so I won't go into further detail. If you have experience with ArcObjects, you can use the esriGeodatabase.IClassSchemaEdit and AlterFieldAliasName methods. I have used this with ArcGIS 9.3 and MS SQL Server but haven't checked to see if these still apply at 10.x. -Steve ... View more

Thanks Jake. MyBad for not looking a little harder. The AddIndex_management tool documentation doesn't mention any of those restrictions and in fact uses a file geodatabase in their example code. -Steve ... View more