How can one change the source data paths for every layer file in folder X using arcpy?
I've followed Updating and fixing data sources with arcpy.mapping as best I can, yet all I get is an unhelpful Runtime error <type 'exceptions.ValueError'>: Layer: Unexpected error
which doesn't tell me enough to troubleshoot what's wrong or missing.
Heres the code (simplified to test a single layer file):
import arcpy, os
fname = r'K:\Layers\xxx.lyr'
lyr = arcpy.mapping.Layer(fname)
oldpath = lyr.workspacePath
print 'oldpath: ', oldpath
lyr.findAndReplaceWorkspacePath(oldpath, r'C:\some\other.gdb')
print 'newpath: ', lyr.workspacePath
and the results:
oldpath: K:\Canvec_Utility\Temp.gdb
Traceback (most recent call last):
File "x10x.py", line 12, in <module>
lyr.findAndReplaceWorkspacePath(oldpath, r'C:\some\other.gdb')
File "C:\ESRI\ArcGIS\Desktop10.0\arcpy\arcpy\utils.py", line 181, in fn_
return fn(*args, **kw)
File "C:\ESRI\ArcGIS\Desktop10.0\arcpy\arcpy\_mapping.py", line 601, in findAndReplaceWorkspacePath
return convertArcObjectToPythonObject(self._arc_object.findAndReplaceWorkspacePath(*gp_fixargs((find_workspace_path, replace_workspace_path, validate), True
)))
ValueError: Layer: Unexpected error
((moved 'update' section into an answer))
Best Answer
It looks like the correct method to use when switching workspaces AND feature datasets is lyr.replaceDataSource(). Here's my working script:
In testing it seems validate in this method is different again: it verifies the new workspace is valid, but ignores the feature class and feature datasets -- meaning it won't return an error if the target FC is not there.
On the other hand if the target FC is present, even inside a different feature dataset, the new data source path is adapted accordingly regardless of whether validate is true or false.
Update: now on Github to allow for easier sharing and revision.