Hi guys. I'm not sure if what I'm observing here is a bug but for sure is something weird. I'm just exporting a table (let's say some million records) from a SDE database (sql server 2008 r2) to a file geodatabase and I get the following strange situations:
When I use ArcCatalog 10.1 (backgroung GP 64 bit) installed on Windows Server 2008 r2 and the file geodatabase is on a local drive the export fails after a while with error 000224: Cannot insert features (which is already weird because I was trying to insert rows !!). When I check the output table I see that some records were exported. If I repeat the same task the failure is the same but at another point in time (different number of records exported before failure)
When I use ArcCatalog 10.1 on Window Server 2008 R2 and the file geodatabase is a on network drive everything is ok
Point 1 and 2 repeated on Windows 7 are just fine. Point 1 and 2 repeated on Windows Server 2008 R2 with 32 bit geoprocessing produce the same trouble.
I'm trying to investigate all the possible cases....meanwhile if anybody has run in a similar trouble.......
After installing pywin for python 2.7 64 bit the problem disappeared probably because the installation triggered some kind of reregistration of the python stuff (just guessing...)
Finally the problem may be identified as follows...
if you install ArcGIS 10.1 with SP1 and background gp 64 bit on windows server 2008 r2 64(but at this point it may be true for windows 7 as well...I did not test it) and you do not install pythonwin you might experience the mentioned problem.
What I said in the previous post is not correct. Installing pythonwin does not fix the trouble at all. I got it wrong. The problem is still there. I would also add that the "table to table" tool does not recognize if the output table already exists in the target file geodb (there is no error message in the form). It works with a folder workspace (saving a dbf) but it does not work for file geodb. This is true for every ArcGIS 10.1 installation I see around here.
If background geoprocessing is disabled and I enter a table name which already exists in the chosen workspace no error message is triggered in the tool form (as it should be). However, error message is produced after running the tool in the notification area (error 000258)
If background geoprocessing is enabled and I enter a table name which already exists in the chosen workspace no error message is triggered in the tool form (as it should be). No error message is produced even when running the tool as it happens for point 1 (back GP disabled). The tool run in any case and overwrites the table
The fact that you can reproduce a similar error means that we are facing a small bug.
As far as the original problem concerns I can confirm you that input table does not seem to be an issue. I tried with different SDE tables ( for example 4 million records with 2 "long" fields, one date and one double), a file gdb table and personal. I would say that the input table should be at least some million records.
The key points are:
The error shows a kind of non deterministic behavior. I start running and after a while it shows the error (cannot insert feature). When I check the output table in the destination ws I see that some records were exported. When I run the tool again the error appears at another point in time because the number of records in the exported table before crashing is different
When the output worlkspace is set using a network path even if I'm pointing to a local drive ("\\machinename\D$\TestGeoDB.gdb") it works
Folder security should not be an issue. Running as administrator and with ConsentPromptBehaviorAdmin set 0 in the registry.
I tried reproducing the original problem, export the Table to a File GDB on local drive, but it completed successfully in my case.
Client- ArcGIS Desktop 10.1 SP1 OS- Windows Server 2008 R2 Standard 64-Bit -- 2GB RAM Database- MS SQL Server 2008 R2 Standard Edition. Table size- Table has approx 2.3 million records (2311211 records to be precise)
I think you are not missing anything. I have sql server 2008 R2 Enterprise but it does not matter because I get the error even if I use another file geodb table as input. Strange. I have this error in all the machines here. I even asked a colleague to make a test and he got the same trouble. There must be a reason linked to my environment here...
Anyway....I can easily workaround about this...the purpose of my ticket was mainly to inform esri of a potential bug. The fact that at least we found a small bug (the overwrite issue) help me not to feel to silly :)
If you want you can close the ticket with ESRI Italy. If I find the reason for that behavior I'll let you know by replying again to this post.