Latest Contributions by KimOllivier

The spatial reference object is read-only. See the old Geoprocessor Programming Model 9.3 PDF. One way to change the central meridian is to manually edit and save a modified spatial reference file after editing in ArcCatalog. Otherwise you have to dump the parameters to a string, do a search and replace and then create an object from the complex string object. I can't believe that this is as intended. import arcpy outputPrj = r'C:\temp\Mollweide (world).prj' spatialRef = arcpy.SpatialReference() spatialRef.createFromFile(outputPrj) stRef = spatialRef.exportToString() stRef = stRef.replace("Central_Meridian',0.0","Central_Meridian',-70.1") spRef2 = arcpy.SpatialReference() spRef2.loadFromString(stRef) print spRef2.name,spRef2.centralMeridian I see that the ArcGIS 10 help suggests some spatial reference properties are read/write. But ArcGIS 9.3 marks them all as read-only. I thought that it may be that sourcing from a protected directory may make them read-only, but that was not the case. I also tried creating from a factory code, setting properties and using sr.create() (which is necessary) but the properties are still read-only. ... View more

There is a good description of programming Word in Python in "Python programming on Win32" by Mark Hammond (O'Reilly) and I found some specific hints in "Writing Excel Macros with VBA" by Roman (O'Reily). Quote from Mark Please note: Word is hard to work with. Here are some examples from his book: http://examples.oreilly.com/9781565926219 ... View more

That is the problem, it does not use the same event queue to detect mouse actions. I don't see anything that you could not do using the tool interface. Have you looked at the validation tab? You can add in Python scripts there that can react to filling in the form before the script is run. You can interactively add in an input there and do some preprocessing depending on the input. ... View more

Its the KISS principle. If all the functions were added that could be done from ArcObjects, then it would be ArcObjects. So the official view would be to use ArcObjects. You may be able to approximate different class breaks by saving the range of breaks you anticipate to be needed, do some statistics and call in the required layer template. If only we had access to an ascii dump of layer objects (as an XML file) we could hack a symbol definition in a layer just like we could with Avenue. Meanwhile it is possible to call ArcObjects from Python using the ctypes module and a good manual on ArcObjects. ... View more

You are not very explicity on what the selected feature is. I assume you mean the page in an index layer in a location map inset? In the index dataframe copy the layer that is being used for data driven pages. The first layer will be an outline say Index. The second layer (renamed say Sheet) will be the highlighted tile. Set this layer to be a filled colour. With data driven pages enabled, open the properties>definition and you will see two button options instead of one, Query Builder or Page Definition. Select the page definition button and choose your options. Set the pages field and set Features that Match. If it is not the index that you want highlighted, the process is the same with any other feature layer, add a field so that the feature(s) to be highlighted has the pagename that corresponds in the attributes. ... View more

Intellisense works just fine in PythonWin while coding if you first import arcpy in the Interactive Window. No need to go to some other IDE. import arcpy will check out a licence but it will make Pythonwin aware of the module. If you only have an arcview licence, then import arcview You can also then browse the arcpy module from Menu>Tools>Browser> arcpy instead of the default __builtins__ However I still find the commentary in the ArcGIS help more useful. ... View more

Deleting layers never worked for me, maybe it does now in 10. My solution is to set gp.overwriteoutput = True or arcpy.env.overwriteOutput = True Then you can always overwrite an in-memory layer definition. ... View more

If you want a more automated approach, you can use the dispatch com interface that we used to use for ArcGIS9. This works well to get direct access to a spreadsheet and each of the sheets. This is well explained in "Python Programming on Win 32" by Hammond and Robinson. o = win32com.client.Dispatch("Excel.Application") o.Visible = 1 o.Workbooks.Add() # for office 97 �?? 95 a bit different! o.Cells(1,1).Value = "Hello" To make accessing Excel a bit easier, Mark Hammond suggested a small module to get the data which I called exceldemos.py: # Excel module from Mark Hammond # Python Programming on Win 32 import win32com.client import win32com.client.dynamic from pywintypes import UnicodeType, TimeType import pprint import os import time import string class easyExcel:     """A utility to make it easier to get at Excel.  Remembering     to save the data is your problem, as is  error handling.     Operates on one workbook at a time."""         def __init__(self, filename=None):         self.xlApp = win32com.client.dynamic.Dispatch('Excel.Application')         if filename:             self.filename = filename             self.xlBook = self.xlApp.Workbooks.Open(filename)         else:             self.xlBook = self.xlApp.Workbooks.Add()             self.filename = ''          def save(self, newfilename=None):         if newfilename:             self.filename = newfilename             self.xlBook.SaveAs(newfilename)         else:             self.xlBook.Save()     def close(self):         self.xlBook.Close(SaveChanges=0)         del self.xlApp        def show(self):         self.xlApp.Visible = 1             def hide(self):         self.xlApp.Visible = 0             def getBooks(self):         return self.xlApp.Worksheets     def getName(self,sheet):         return self.xlBook.Worksheets(sheet).Name # #    now for the helper methods #     def getCell(self, sheet, row, col):         "Get value of one cell"         sht = self.xlBook.Worksheets(sheet)         return sht.Cells(row, col).Value         def setCell(self, sheet, row, col, value):         "set value of one cell"         sht = self.xlBook.Worksheets(sheet)         sht.Cells(row, col).Value = value         def getRange(self, sheet, row1, col1, row2, col2):         "return a 2d array (i.e. tuple of tuples)"         sht = self.xlBook.Worksheets(sheet)         return sht.Range(sht.Cells(row1, col1), sht.Cells(row2, col2)).Value         def setRange(self, sheet, leftCol, topRow, data):         """insert a 2d array starting at given location.         Works out the size needed for itself"""                 bottomRow = topRow + len(data) - 1         rightCol = leftCol + len(data[0]) - 1         sht = self.xlBook.Worksheets(sheet)         sht.Range(             sht.Cells(topRow, leftCol),             sht.Cells(bottomRow, rightCol)             ).Value = data     def getContiguousRange(self, sheet, row, col):         """Tracks down and across from top left cell until it         encounters blank cells; returns the non-blank range.         Looks at first row and column; blanks at bottom or right         are OK and return None witin the array"""                 sht = self.xlBook.Worksheets(sheet)                 # find the bottom row         bottom = row         while sht.Cells(bottom + 1, col).Value not in [None, '']:             bottom = bottom + 1                 # right column         right = col         while sht.Cells(row, right + 1).Value not in [None, '']:             right = right + 1                 return sht.Range(sht.Cells(row, col), sht.Cells(bottom, right)).Value     def getHeaderRange(self, sheet, row, col):         """Tracks  across from left cell until it         encounters blank cells; returns the non-blank range.         Looks at first row and column; blanks at bottom or right         are OK and return None within the array"""                 sht = self.xlBook.Worksheets(sheet)                 # find the bottom row         # bottom = row         # while sht.Cells(bottom + 1, col).Value not in [None, '']:         #     bottom = bottom + 1                 # right column         right = col         while sht.Cells(row, right + 1).Value not in [None, '']:             right = right + 1                 return sht.Range(sht.Cells(row, col), sht.Cells(row, right)).Value        def fixStringsAndDates(self, aMatrix):         # converts all unicode strings and times         newmatrix = []         for row in aMatrix:             newrow = []             for cell in row:                 if type(cell) is UnicodeType:                     newrow.append(str(cell))                 elif type(cell) is TimeType:                     newrow.append(int(cell))                 else:                     newrow.append(cell)             newmatrix.append(tuple(newrow))         return newmatrix             def test():     # puts things in a new sheet which it does not save     spr = easyExcel()     spr.show()         input = 'hello'     spr.setCell('Sheet1',1,4, input)     output = spr.getCell('Sheet1',1,4)     print output     assert input == output, 'setCell/getCell failed'         input = []     for i in range(10):         row = []         for j in range(4):             row.append(str('(%d,%d)'% (j, i)))         input.append(tuple(row))         spr.setRange('Sheet1',2,2,input)         output = spr.getRange('Sheet1',2,2,11,5)     # get rid of unicode strings     output = spr.fixStringsAndDates(output)     assert input == output, 'setRange/getRange test failed'         #get a contiguous range     output2 = spr.getContiguousRange('Sheet1',2,2)     dimensions = (len(output2), len(output2[0]))     assert dimensions == (10, 4), 'getContiguousRange failed'         print 'passed!'     # if __name__ == '__main__' : #     test()     Then to get each sheet you can create a Python loop something line this: The spreadsheet had data split into several tables, with headers. I needed to transpose the data and calculate some cross sections from a centreline. In my case I exported to CSV files and loaded them using another obsolete tool called CreateFeaturesFromText, but you could write the data directly into a table using a cursor. # load reachs and crosssections from a spreadsheet # getwater3.py # get channel centreline and number nodes uniquely # dropping off duplicates # calculate bearing of nodes in geographic degrees # Kimo # 6 October 2005 # upgrade to average angles for middle points of a reach # 7 October 2005 # 10 October 2005 # offset one more row to allow for max water level calc # does not get all the other water levels, just bottom and max # better to load database of levels for each section for joining later # so that we can create a surface for each time step from same cross sections # ref first field for sorting and joining # pad chainage # removed max row again # but too late, have to have it. # # import win32com.client import win32com.client.dynamic import os,sys from math import * # sys.path.append("d:\\project\\python") import exceldemos homedir = "e:/toolbox/scratch" os.chdir(homedir) xlApp = win32com.client.dynamic.Dispatch('Excel.Application') filename = "e:/example/GISTransfer.xls" spr = exceldemos.easyExcel(filename) f1 = open("channel.txt","w")     # polyline generate file f2 = open("channel_att.txt","w") # polyline attributes f3 = open("channel_node.txt","w")# xsection nodes, can load as XY events f4 = open("channel_skip.txt","w")# rejects - duplicate nodes f5 = open("channel_xsec.txt","w")# xsection ref,x,y,theta for joining to xsections # there are multiple worksheets # so loop over all to combine into single output max = 0 lstRef = [] try :     for b in spr.getBooks() :         max += 1         print "workbook: ",b.Name     print "Workbooks found: ",max     row = 0     # output file headers     print >> f1,'Polyline'   # for Toolbox loader     print >> f2,'rid,reach'     print >> f3,'ref,nodeid,x,y,z,w,theta,rid,reach,chain'     print >> f4,'rec,x,y,reach,chain,z,w'     # no header for f5 xsec     # print >> f5,'ref,x,y,theta,maxwater'     rec = 1     id = 0     rn = 0     # skip first sheet     for sheetindex in range(2,5) :                 # get number of columns         header = spr.getHeaderRange(sheetindex,3,1)         columns = len(header[0])         print "Length of header",columns                 print "Beginning :",spr.getName(sheetindex)         aHeader = spr.getRange(sheetindex,1,1,6,columns)         # transpose rows and columns         headerpivot =  [[r[col] for r in aHeader] for col in range(len(aHeader[0]))]         # initialize counters         rname = ""         rx = 0         ry = 0         i = 0         rej = 0         colcount = len(headerpivot[1:])         print "Columns",colcount         for (a,b,c,d,e,w) in headerpivot[1:] :               # print rec,a,b,c,d,e,w             # a X coordinate             # b Y coordinate # ..... snip to fit forum text limit .....     print >> f1,"END"     print "records :",rec,"nodes",id,"reaches",rn,"rejects",rej     f1.close()     f2.close()     f3.close()     f4.close()     f5.close()     spr.close()             del xlApp except :     print "cleanup close"     f1.close()     f2.close()     f3.close()     f4.close()     f5.close()     spr.close()         del xlApp     The spreadsheet is a bit large to include, I hope this is enough of a clue to encourage you to buy the book, which is now available in digital format. http://oreilly.com/catalog/9781565926219 ... View more

It is because the paths to Python26 are not set for Arcpy. Even if you reinstall Python26 you have to make sure there is a Desktop10.pth file in c:/Python26/Lib/site-packages containing the paths to the package C:\Program Files (x86)\ArcGIS\Desktop10.0\bin C:\Program Files (x86)\ArcGIS\Desktop10.0\arcpy C:\Program Files (x86)\ArcGIS\Desktop10.0\ArcToolbox\Scripts Also you might need to reinstall Numpy which resides in site-packages as a folder Note that if you use the ArcGIS installer it will install another copy of python in C:/Python26/Desktop10.0/Lib/.... so if you cleanup and reinstall Python yourself the configuration will be broken. But on the other hand if you use the Esri installer, the Python install will be broken, you will not be able to open the Python help and other tools will fail. ... View more

The "Text loader", part of Microsoft libraries needs a Schema.ini to properly define the field names and types. ArcGIS uses it. If schema.ini does not exist it defaults to using the header fields which will then fail if they are not valid field names. With a schema.ini you can rename the fields on import and also define the field type. The schema.ini is a recent discovery for me, and has saved me hours of work creating custom scripts. It is mentioned in the ArcGIS help, but do a Google search for more details from Microsoft. http://msdn.microsoft.com/en-us/library/ms709353%28v=vs.85%29.aspx ... View more

Renaming a layer in the current MXD should be instantaneous. After all the MXD will be in memory. This took 312 milliseconds. I think that having things on a network will be the likely culprit. It could be a very slow licence manager lookup causing the delay. Try a ping and other network analysis tools. I am amazed at the lengths people go to to avoid a simple local machine running Windows with local disks. They surely are a commodity? This is the only way I run processes for GIS analysis. Anything else is asking for much delay, even network calls are not optimised in any system, they simulate a local disk call, but are very slow across a network. Nothing to do with the network drives, its the OS and the network disconnection. Here is a benchmark script. I ran this in the Python window. >>> # rename layer ... import arcpy ... import datetime ... start = datetime.datetime.now() ... mxd = arcpy.mapping.MapDocument("CURRENT") ... lstDF = arcpy.mapping.ListDataFrames(mxd) ... print len(lstDF),"frames" ... for df in lstDF: ...     print df.name ...     llay = arcpy.mapping.ListLayers(mxd,"",df) ...     for lay in llay: ...         print lay.name ...         print mxd.title,df.name,lay.name ...         lay.name = lay.name+"changed" ... arcpy.RefreshActiveView() ... arcpy.RefreshTOC() ... print datetime.datetime.now() - start ... 2   frames Layers pointerchanged topo250   Layers   pointerchanged coastlinechanged topo250   Layers   coastlinechanged lakechanged topo250   Layers   lakechanged native_polychanged topo250   Layers   native_polychanged island_polychanged topo250   Layers   island_polychanged mainlandchanged topo250   Layers   mainlandchanged index250changed topo250   Layers   index250changed Index coastlinechanged topo250   Index   coastlinechanged sheetchanged topo250   Index   sheetchanged index250changed topo250   Index   index250changed 0:00:00.312000 ... View more

I started a similar project to create a tool that split a parcel in half if there were two address points in the parcel, when the parcel was a corner property. It needed ET_Geotools I created a line for the split by bisecting the two points and used that to intersect the polygon. You can then write the split polygon back to the featureclass. # bisect2 # bisect a parcel into two # between two points inside # extend to multipoints later # input parcel_id # select parcel poly # select address points # if two points # calculate perpendicular bisector # cut polygon with the line # save to be replaced # 26 Nov 2010 # import arcgisscripting,sys,os,math gp = arcgisscripting.create(9.3) def createknife(par_id):     """create knife lines for bulk splitting     using address points with same parcel_id"""     gp.MakeFeatureLayer_management(fcAdd,"add_lay","parcel_id = "+str(par_id))     shapeField = gp.Describe(fcAdd).shapeFieldName     cur = gp.SearchCursor("add_lay","parcel_id = "+str(par_id),SR)     row = cur.next()     n = 0     pt = []     while row:         pt.append(row.getValue(shapeField).centroid)         n+=1         row = cur.next()     del row,cur     ## print n, "points"     ##for addpt in pt:     ##    print "address ",addpt.X,addpt.Y      # find mid point and angle, add 90 degrees     midpt = gp.CreateObject("Point")     midpt.X = (pt[0].X + pt[1].X)/2.0     midpt.Y = (pt[0].Y + pt[1].Y)/2.0     radAngle = math.atan2(pt[0].Y - pt[1].Y,pt[0].X - pt[1].X) + math.pi/2.0     # print "midpoint",midpt.X,midpt.Y,"Bearing",radAngle     # create a perpendicular line in a temp fc     ext = gp.Describe("in_memory/parcel").extent     size = max(ext.width,ext.height)*1.5     print par_id,size     pnt1 = gp.CreateObject("Point")     pnt2 = gp.CreateObject("Point")     pnt1.X = midpt.X - size * math.cos(radAngle)     pnt1.Y = midpt.Y - size * math.sin(radAngle)     pnt2.X = midpt.X + size * math.cos(radAngle)     pnt2.Y = midpt.Y + size * math.sin(radAngle)     pline = gp.CreateObject("Array")     pline.Add(pnt1)     pline.Add(pnt2)     ##print "knife    %8.1f %8.1f %8.1f %8.1f" % (pnt1.x,pnt1.y,pnt2.x,pnt2.y)     cur = gp.InsertCursor("in_memory/knife")     row = cur.newRow()     row.shape = pline     cur.insertRow(row)     row = cur.newRow()     pline.removeAll()     del cur,row     # temp debugging     outLine = gp.CreateScratchName("split_line","","FeatureClass")     gp.CopyFeatures_management("in_memory/knife",outLine)     ##print "Knife   extent",gp.Describe("in_memory/temp").extent     ##print "Parcel  extent",gp.Describe("in_memory/parcel").extent     ##print "Address extent",gp.Describe("add_lay").extent     ##outAdd = gp.CreateScratchName("split_add","","FeatureClass")     ##gp.CopyFeatures_management("add_lay",outAdd)     gp.AddToolbox(r"C:\Program Files\ET SpatialTechniques\ET GeoWizards 10.0 for ArcGIS 9.2 and 9.3\ET GeoWizards.tbx")     outParcel = "in_memory/split" #gp.CreateScratchName("split_parcel","","FeatureClass")     gp.ET_GPPartitionPolygons("in_memory/parcel","in_memory/temp",outParcel)     # print gp.GetMessages()     return outParcel # -------------------- main ----------------------- try:     fcPar = sys.argv[1] except:     fcPar = "corner" gp.overwriteOutput = True ws = "d:/work/pcl/CourierRun.gdb" gp.workspace = ws fcAdd = "addrun" SR = gp.CreateObject("SpatialReference") SR.CreateFromFile("c:/arcgis/nztm.prj") gp.CreateFeatureClass("in_memory","knife","Polyline","","","",SR) rCursor = gp.SearchCursor(fcPar) row = rCursor.next() while row:     par_id = row.par_id     result = split(par_id)         row = rCursor.next() ... View more

1. Load the spreadsheet into a table. If you dump the spreadsheet to a .CSV file first and define a Schema.ini then it will load nicely as a formatted table using the CopyTable_management tool. Schema.ini is a default Windows function. Note that a spreadsheet is not a database and is not supported to be reliably loaded with a satisfactory schema because columns are not defined, only formatted. Default widths are 255 chars for example. 2. If you have to update multiple sheets from one table, create a dictionary using the key value and the attribute. 3. Run an updateCursor on each target table and push the new values in using the dictionary as a lookup. Using the join tools are a bit slow. ... View more

To debug a tool, try this: Run a tool interactively by filling in the dialog by hand. After it runs successfully, open the results and right-click to get "Copy as a Python Snippet" Paste this into an empty Python script and examine the syntax, particularly the SQL query. I find that provided you are not using a personal geodatabase (*.mdb) AND you are using valid field names (No spaces or silly chars, starts with a letter, less than max chars, not a reserved word) you can also dispense with the double quotes around the field names. In spite of the help saying otherwise. This makes assembling SQL queries much simpler. Also try printing out your expression to see if you really got the expected valid expression. ... View more

I used the standard module ElementTree to parse XML for a GPX file as hinted by a Python demo at the conference. This is much easier than some of the older parsers such as SAX or DOM. ... View more