Latest Contributions by KimOllivier

I've actually tried VB and Python (and every punctuation mark I could think of) and neither works. I ran out of things to try and ended up just doing this in ArcMap. I think the Calculate_Field tool is buggy or something because it works sometimes and othertimes it doesn't. It is more likely that you have data that causes a crash in the expression. If you use the field calculator you do not have the opportunity to trap data errors in a try/except block and its impossible to see what record caused it. Since the CalculateField simply wraps the expression in a cursor, it is much much better to use your own cursor in a script and do the expression inside a try/except block. CalculateField is more useful in interactive processes or models. ... View more

fperks;16230 wrote: Heres an easy way to make arcpy non iterable objects usable in a for each loop. for each in iter(lambda: <your object>.next(), None): This basically iterates through getting the next element until the termination condition of the element being None is met. Thanks very much. Wonderfully Pythonic! ... View more

I haven't really seen any bugs regarding comtypes and ArcObjects, except one:  it doesn't wrap the esriSystem.Array object since there's a name clash with ctypes Array.  I'd be curious as to what version of comtypes you've been using and what problems you've been seeing. I did notice that in arcpy the Array object is not iterable. I had to use the .next() function. Should I expect it to be iterable? while row:         feat = row.getValue(shapefield)         qInterior = False         for partNum in range(feat.partCount) :             part = feat.getPart(partNum)             qInterior = False             for ptNum in range(part.count):                 pt = part.next()                 if pt != None:                     arrayOuter.add(pt)                 else :                     qInterior = True                     break # ignore donut vertices             arrayObj.add(arrayOuter)             arrayOuter.RemoveAll() ... View more

Yes, its a real pain! Thefull  list is in a couple of  PDF files  buried in the documentation folder in the ArcGIS install location. I have a little text file on my desktop that I can open and cut-and-paste the required transform string. ... View more

I see that in one context the domain is included in the spatial reference object. However it is not included in the spatial reference projection file. So I was looking in the wrong place. Unless I looked up the geoprocessing diagram I would not have noticed that it was included. The issue of gp.Merge creating a different domain is still a problem. We are not supposed to have to set these with high precision geodatabases any more. Surely that is a bug? ... View more

ArcGIS 10 puts its own copy of Python26 in a subdirectory of Python26/ArcGIS10.0 (!) When you install your own Python26 it does not copy the link for ArcGIS python extensions move Desktop10.pth and the numpy folder back to C:/Python26/Lib/site-packages and all will be standard and well. You can then delete the whole ArcGIS10.0 folder so that you only have one version of 2.6 installed ... View more

You can define an output projection (and transform) for a cursor and write out projected coordinates. # extent_xy.py # add extents of parcels in NZMG to a NZTM dataset for backward compatibility # update to use on the fly projection directly from NZTM layers # see Werner Flacke's book and script example # updated for 9.2 # Kim Ollivier 13 April 2007 # input NZTM coverage library, parcel and plabel coverages # output csv files to load and join back to coverages using AML # issues solved: #               add spatial reference to SearchCursor #               define a Geographic Transformation that works with a coverage #               because coverages cannot define NZGD2000 use WGS84 definition #               or the transform is ignored import os,sys,glob,arcgisscripting gp = arcgisscripting.create() def parcel(tile) :     print "Parcel",tile     ws = "e:/lib/nztm/tile/t"+str(tile)+"/data"     outfolder = "e:/lib/nztm/tile/t"+str(tile)+"/data"     ds = ws + "/parcel"     f1 = open(outfolder+"/e_parcel.txt","w")     f1.write("par_id,eminx_nzmg,eminy_nzmg,emaxx_nzmg,emaxy_nzmg\n")     gp.Workspace = ds     print ds     # this is a good on-the-fly switch for the searchcursor     # srOut must be a com object, not a file ref or a factory code     rows = gp.SearchCursor(ds+"/polygon",'"PAR_ID" > 0',srOut)     n = 0     rows.Reset()     row = rows.Next()     while row  :         print >> f1,"%d,%s" % (row.par_id,row.shape.Extent.replace(" ",","))         row = rows.Next()         n +=1     del rows     print n,"Polygons"     f1.close()     return def plabel(tile) :     print "Plabel",tile     ws = "e:/lib/nztm/tile/t"+str(tile)+"/data"     ds = ws + "/plabel"     outfolder = "e:/lib/nztm/tile/t"+str(tile)+"/data"     f2 = open(outfolder+"/e_plabel.txt","w")     f2.write("par_id,eminx_nzmg,eminy_nzmg\n")     gp.Workspace = ds     rows = gp.SearchCursor(ds+"/point",'"PAR_ID" > 0',srOut)     n = 0     rows.Reset()     row = rows.Next()     while row  :         n += 1         print >>f2,"%d,%s,%s" % (row.par_id,row.shape.Extent.split()[0],row.shape.Extent.split()[1])         row = rows.Next()     del rows     print n,"plabels"     f2.close()     return # ------ main ---- # srOut = gp.CreateObject("SpatialReference") srOut.CreateFromFile("c:/arcgis/nzmg.prj") # WGS_1984 works for coverages defined with an equivalent custom Transverse projection definition gp.GeographicTransformations = 'NZGD_1949_To_WGS_1984_3_NTv2' # ;New_Zealand_1949_To_NZGD_2000_3_NTv2' # but is this being used? inFC = "e:/lib/nztm/tile/t1009/data/plabel/point" # not at 9.3?? outFC = "in_memory/dummy1.shp" outFC = "c:/tmp/dummy1.shp" outPrj = "c:/arcgis/nzmg.prj" # cannot be a COM object or factory id 43040 geoTrans = "NZGD_1949_To_WGS_1984_3_NTv2" # ;New_Zealand_1949_To_NZGD_2000_3_NTv2" if gp.Exists("c:/tmp/dummy1.shp") :     gp.Delete("c:/tmp/dummy1.shp") # gp.OverwriteOutput = 1 gp.MakeFeatureLayer_management(inFC,"tlayer","PLABEL# < 3") gp.Project_management("tlayer",outFC,outPrj,geoTrans) try :     if sys.argv[1].upper() == 'ALL' :         lstTile = range(1001,1013)     else :         lstTile = [ tile for t in sys.argv[1].split(",")] except :     lstTile = range(1001,1013) print "begin extent_xy"    for t in lstTile :     print t     parcel(t)     plabel(t) ... View more

I think I can guess what is happening. You will get into trouble deleting records while looping through the same set. The only way to make this safe to delete records is to make a copy of the keys or the full set and use that in the loop. ... View more

Yes, this is a very successful way of creating an SQL query that is efficient. A keyfile type selection is often wanted, and the MakeQueryTable is buggy at 9.3 and has a lot of limitations. You have to have all features in the same geodatabase, complex expressions are not supported in file geodatabase and so on. Here is my workaround: Loop through the first table with a cursor to collect a list of keys. Pass it through a SET structure to remove duplicates. Create a (potentially huge) sql expression Use this expression on the source featureclass to create a layer Process or export the layer to a new featureclass # to make a keyfile # 26 April 2010 kimo # much better than MakeQueryTable and it works across geodatabases # still uses an SQL query but python generates the set for the expression # will work for tens of thousands of keys (amazing) # if you want to save the featureclass or table as a layer or view definition # to use in ArcMap (without making a copy) you MUST have an index on the key field # or it will be extremely slow to draw. # see the Keyfile tool in ArcScripts for an alternative method using selection sets in ArcMap import arcgisscripting,sys,os,datetime,fileinput gp = arcgisscripting.create(9.3) def titlekey(title) :     gp.Workspace = ws     # Make a title set to use as a keyfile     lstTitle = []     cur = gp.SearchCursor(title)     row = cur.next()     while row :         lstTitle.append(row.ttl_title_no)         row = cur.next()     del row,cur     return set(lstTitle) def title_tab(src) :     " use setTitle to export records to a table"     start = datetime.datetime.now()     gp.Workspace = ws     setTitle = titlekey(title)     expTitle = ",".join(["'"+t+"'" for t in setTitle])     sqlQuery = "ttl_title_no in ("+expTitle+")"     gp.MakeTableView_management(src,src+"_view",sqlQuery,wsout)     gp.OverwriteOutput = True     gp.CopyRows_management(src+"_view",wsout+"/"+src+"_sub")     print gp.GetCount(wsout+"/"+src+"_sub").Getoutput(0),"records in "+src+"_sub"     print datetime.datetime.now() - start # ------------------------- main ------------------------- title = "e:/arrears.mdb/title" ws = "e:/memo.gdb" wsout = "e:/arrears.mdb" # create subsets based on title keys if not gp.Exists(wsout+"/tta_sub") :     title_tab("tta") ... View more

Not yet in Prerelease 10. Even the 9.3 pdf has been removed, which I think is a mistake, since we will need to support a lot of old style scripts for years. Keep your own copy before uninstalling 9.3! I hope that there will be a new diagram. Not holding out hope for a new book. The authors have moved on to other projects. ... View more

I have an icon on the start menu at Programs>Python 2.6>Pythonwin. Of course I had to hack the Explorer filetypes to set the edit paths properly for .py files, but I have to do that in every training course since 9.1 for all the students. I am still annoyed at the install hijacking the Python 2.6 install directory with another level under C:\Python26\ArcGIS9.4\ .  Yuck! a full stop in a directory path? How non-standard is that? This apparently does not happen if Python 2.6 is already installed, so perhaps Python should be installed before ArcGIS, as a prerequisite just like .NET. Then apparently we get ArcGIS to respect the normal structure. I will give it a try to see if I can defeat the odd behaviour at Beta 2 by installing Python again first. ... View more

Hope you might look at these procedures in the future to make contributions from over seas a bit easier... At least for beta testers. Surely the beta team want direct feedback? I am sorry that reporting through the distributor support desk is not worth the effort for me so I just ignore the requests to report a beta bug that way. We know there will be bugs, its not a production issue. If the beta team are not monitoring the beta feedback forums then why bother at all? The agreement requires active testing and feedback. ... View more

I also use mapbook extensively to generate atlases and also to render raster images for an image catalog. This is still hard to get right, I have to revert to page imperial units to match dots-per-inch exactly to get the right number of pixels in a page with no borders etc. An option to export the world file of the page when exporting as an image would save me another painful calculation step in Python, using an index, finding the corners of the page and doing my own transform. When you get down to the details to look at the edges of tiles exported like this  it is clear that the X and Y scales are slightly off, so the tiles are not seamless. So yet another external package adjust to rescale, resize and reclip to extract out an accuractly scaled and registered image. ... View more