Latest Contributions by TimLangner

Not only that but one can also create all of the following as well: 1. Slef intersections 2. incorrect ring ordering 3. incorrect segment orientation The last two are a bug: NIM074901. I still have an on going support call on the matter so things may change but current thinking is that the first one is caused by user digitising. In my case I believe it comes about from the way ESRI works when you insert points and the fact inserted points cannot be automatically snapped to an existing point when inserted. It also relates to the set tollorance levels. I left the levels to their previous setting when some months ago I imported the data. They are currently set to 0.0001. Apparently at face value the points appear to be snapped but when you check the topology they do not match. So are obviously very close to the 0.0001. The suggestion was to use the Check Geometry as a warning of poor, although not strictly incorrect, digitising and correct the defect. Of course that isn't allowed on the SDE and as my dataset has an archive, it can't be exported, since no correct export options exists for an archive. I can export parts of the archive but not the whole archive. I'm using ArcMap 10.0.3 and yet Redlands ended up referred me to the following link: http://support.esri.com/en/knowledgebase/techarticles/detail/29465 This just goes to prove that issues exist in this area because that article refers to being able to check geometries and it states software up to 9.0. It also says for ArcGIS 9.1 and later, the Check Geometry tool is disabled for ArcSDE data. It may be a case of ESRI wishing to introduce a feature and deciding another feature will have to be sracped so it can be introduced, rather than finding away of inluding both features. I keep finding things that can't be done in ArcMap that I'd like to do. It's very good but it could be even better, ... View more

Possibly the reason edges do not match as reported in other parts of  this thread is that densification may be occurring in other  geoprocessing steps and it will always occur if the output is to a  Shapefile. My data was held via an ArcSDE Geodatabase and stored in Oracle 10g database. In one case I was only digitising; clipping; spiting and reshaping using one feature class. It would be a case of creating a feature; creating another and then clipping the larger feature to create a whole or a mixture of this. I might also do some tracing. More recently I raised the issue with ESRI UK, as I found geometric errors in my ArcSDE geodatabase feature class and I couldn't do a dissolve. By being able providing clear steps to them in respect of the incorrect ring ordering error, they were able to replicate the issue very quickly. It since has been given the bug number: NIM074901 [http://support.esri.com/en/bugs/nimbus/TklNMDc0OTAx]. I still have outstanding issues with respect to incorrect segment orientation and self intersections. With respect to the latter, an ESRI support guy checked and clean up one particular self intersection, after exporting an extract of my data to a file geodatabase. This then deleted a single point. However when looking at it, he couldn't see that the point, that was deleted, was self intersecting. It is in the early stages with that investigation so it may still be intersecting, although it appears not to be. These and the other errors were all created whilst digitising; clipping; splitting or  reshaping the same feature class. Some of the reshaping was against data imported from shapefiles and I was tracing this existing data from time to time. However was all stored in an ArcSDE geodatabase at the stage the errors had occurred. I feel all the issues are possibly caused during the translation of ESRI format data to Oracle's SDO_geometry format. I've also found ways to generate null geometries. This is related more to working with the attribute data but it also shows that invalid geometries can be entered in the an ArcSDE Geodatabase. Personally I think the best issue at the moment would be to offer a repair geometry function for data stored in such a geodatabase. It's certainly very interesting problem. Kind regards ... View more

ESRI Inc. in America have issued a bug number for this problem. It is NIM074901. I'm not sure if this covers the self intersections geometry errors I was getting or just the incorrect ring ordering. I'm stil in contact with ESRI(UK) Inc. so if I get anything else I will post it here. Kind regards Tim ... View more

Thank you for your comments. That is just what I was looking for. I assume could I then create a view of the results which is accessible in ArcMap, although it wouldn't be possible to update it but then it isn't possible to do that is a Make Query Table, which I have been using. I actually did the distinct work in Excel this time and then used the Make Query Table to join this new table. I then exported this new table because it isn't possible to do a relate on a Make Query Table. For sound reasons I am sure but then I fail to see the sound reasons why a make query table can't be edited in the first place. Hopefully ESRI will add such a feature in a later release of ArcMap. By importing this new table I could do a relate to get the rows that did not match and I could then manually update them. Quite long winded but it at least allowed me to do what I required. Kind regards Tim ... View more

Looks like I am going to have to experiment on this one. Hi Raymond. Did you ever find a work around? I spoke to ESRI (UK) about this issue and they confirmed that it is not possible to create a duplicate of an archive. Using feature class to feature class causes it to become orphaned. I've not investigated it fully to see if it is the archive which is orphaned and whether ti creates many orphaned feature classes if many changes have been made over time. I put up another thread here: http://forums.arcgis.com/threads/40062-History-of-versioned-data-after-feature-class-to-feature-class?highlight=backup and an ArcGIS idea here, which can be promoted: http://ideas.arcgis.com/ideaView?id=08730000000brpcAAA Kind regards Tim ... View more

Hi Brent No what I want is a complete copy of the whole archive. Not  any set point in time but the whole archive. Imagine I was going to  publish a book detailing the chronological history of a place. Say I give the  book to the printers to print 10,000 copies. The printer wouldn't just print the  final section containing the most recent history, no they would print every page  in the book. That is what I wish to achieve when doing a feature class to  feature class copy. Although our complete geodatabase gets backed up, it is  backed up by our IT department and they have no GIS knowledge or knowledge of  ArcSDE. Thus it is far easier for me to restore the feature class from my own  backup copy, should any data issues arise, then asking IT to do it. However I  cannot do this with archiving enabled. I have added a post to ESRI ideas.  So if anyone else would like this extremely useful function, add your support  there. Not signed into or don't have an ESRI account use this link: http://ideas.arcgis.com/ideaView?id=08730000000brpcAAA Signed into an ESRI account use this link: https://c.na1.visual.force.com/apex/ideaView?id=08730000000brpc&returnUrl=%2Fapex%2FideaList%3Fc%3D09a300000004xET%26sort%3Drecent Kind  regards Tim ... View more

A while ago I recall hearing the built in select by attributes dialogue having already started querying the data when it is opened (list of fields, names, joins etc) which already takes care of some of the overhead that the arcpy.SelectLayerByAttribute_management() still has to go through. Making different selections based on changing criteria that you input will not gain you any performance using python scripts than using the built in tools, or else Esri would have coded the built in dialogue to be the same as the arcpy tool. The performance gain is in automation. Not having to query the same (or similar) data over and over and over and over again. Another method you may want to try is creating a feature layer query. http://help.arcgis.com/en/arcgisdesktop/10.0/help/index.html#//00170000006p000000 Using a where clause with this tool may give you the performance speed you are looking for, depending on why you need to make a selection at all. Thanks for your reply Matt. What I'm doing is a series of quick visual and data checks. 1. I am checking that features are assigned to the correct road and using the correct UPRN (UK address id). 2. If the features are assigned to the wrong road then I need to switch to the LLPG layer in my query and search on the street name so that I can then update the details of the incorrect assigned features. It's a fair quick task in terms of the query itself but requires me to switch between the two layers constantly. The queries are very similar because the tables use the same field names for the fields I am checking. However every time I switch layers in ArcGIS, in the select attributes window, it removes the what I have typed. In addition to that it is very easy to lose the Select attribute window, as it gets hidden behind the create feature window; table and identify windows which I have open. If I then go to the menu and hit Select by attributes, nothing appears and I then have it hunt around to find it hidden behind other windows. I could put it on the the map side but it obscures the map. I'd rather leave it open than close it. By using the Python window I can use ALT+TAB to switch to it. I can amend my script,save and hit run. It's not possible to switch however to the select by attributes window in ArcGIS. I guess my issues are really myself trying to find ways around annoying GUI interface issues, as opposed to a genuine need to do this in python but I thought perhaps Python might be a good workaround, until ESRI, if they ever do so, improve the GUI interface. What would be good is something along the lines of SQL Developer. In fact you can have multiple tabs open and from those run multiple queries and scripts which all output to the output window. Kind regards Tim ... View more

Hi Matt Your quite right. It was that. I was using the syntax for personal geodatabases when I'm working with an SDE geodatabase. My revised code is: arcpy.SelectLayerByAttribute_management ("Full area polygons layer outline only v5", "NEW_SELECTION", """ [AREA_NAME] = 'AREA 3' """) Saying that after I got my new code to work and it had selected the features, ArcMap shortly froze. However on restarting ArcMap it then worked, after several error messages about valid index range being passed. That may be a freak event, as I'm new to working with the Python window. Kind regards ... View more

Hi Stancy, Thank you for your reply. I solved the problem in the end. In an exception statement I can have multiple if and else statements, which I can use to trap my exceptions and call the procedures that I have written to handle them. Thus I don't need to create my own exception classes. This is a simpler solution and works for the time scale I have. May be one day I'll have time to get my head about creating my own exceptions. Kind regards Tim ... View more