topic Re: Alternative for Wildcards in Update Cursor in Python Questions

Alternative for Wildcards in Update Cursor

DonDailey — Sun, 12 Dec 2021 06:28:59 GMT

I've got a point feature class of road intersections in which I used a script to create a field of joined road names from a line feature class. I've got to separate each road name into separate fields. I've been able to do that without much problem. But the road names were joined in more or less random fashion and now I need to rank the road names based on their type into their own field i.e. Interstates in the first field, US Highways in the second, state highways in the third, etc. All the highways begin with either Interstate, US or State.

I've tried doing a wildcard if ... else on the list created by the .split inside the update cursor but apparently that's not allowed. From searching it seems that regular expressions (which I've never used) might be my only hope. Anybody have any suggestions?

The code I used to split the field and place the road names into individual fields.

try:
# Import arcpy module
    from arcpy import env

#set workspace
    env.workspace = r"GDB"

# Local variables:
    fc = r"feature class"
    
#create update cursor to place values in fields
    fieldList = ["AllRoutes","ID_Road1_Cross","ID_Road2_Cross","ID_Road3_Cross",
                 "ID_Road4_Cross","ID_Road5_Cross"]
    with arcpy.da.UpdateCursor(fc,fieldList) as cursor:
        for row in cursor:
              cleaned = [x.strip() for x in row[0].split("&")]
              if len(cleaned) == 2:
                row[1] = cleaned[0]
                row[2] = cleaned[1]
              elif len(cleaned) == 3:
                row[1] = cleaned[0]
                row[2] = cleaned[1]
                row[3] = cleaned[2]
              elif len(cleaned) == 4:
                row[1] = cleaned[0]
                row[2] = cleaned[1]
                row[3] = cleaned[2]
                row[4] = cleaned[3]
              elif len(cleaned) == 5:
                row[1] = cleaned[0]
                row[2] = cleaned[1]
                row[3] = cleaned[2]
                row[4] = cleaned[3]
                row[5] = cleaned[4]
              cursor.updateRow(row)


except arcpy.ExecuteError:
    print arcpy.GetMessage(2)

except:
    print "Process did not complete"

Re: Alternative for Wildcards in Update Cursor

DanPatterson_Retired — Wed, 20 Apr 2016 22:08:05 GMT

I formatted your code Code Formatting... the basics++

Re: Alternative for Wildcards in Update Cursor

DarrenWiens2 — Sun, 12 Dec 2021 06:29:01 GMT

It's getting to the end of a long day and I'm having a hard time deciphering the problem on this one, so bear with me. If you want to know if a string starts with something, you can use the startswith() method:

if myString.startswith('Interstate'):
  return 'Interstate'

Re: Alternative for Wildcards in Update Cursor

curtvprice — Sun, 12 Dec 2021 06:29:04 GMT

What happens when two interstates intersect? Can Field1 and Field2 both be interstates?

If so I suggest (inside the cursor with block) assigning a key and sorting, like this:

cleaned = [x.strip() for x in row[0].split("&")]  

rds = []
for rd in cleaned:
    rdup = rd.upper()
    if rdup[0] == "I": key = 0  # Interstate
    elif rdup[0] == "U": key = 1 # US route
    else: key = 2 # state and other
    rds.append([key, rd])
# sort and place in fields
for k, rd in enumerate(sorted(rds)): 
    row[k+1] = rd[1] # k = 0,1,2..
rows.updateRow(row)

Re: Alternative for Wildcards in Update Cursor

DanPatterson_Retired — Thu, 21 Apr 2016 01:17:58 GMT

Just curious... how many records are there in the road dataset to begin with?

Re: Alternative for Wildcards in Update Cursor

DonDailey — Thu, 21 Apr 2016 01:40:59 GMT

I'm testing this out for a single rural county that has about 1,290 intersections on all public roads. I'm sure to do all counties it will be over 100,000 records. Right now I'm just trying to prove it can be done.

Re: Alternative for Wildcards in Update Cursor

DanPatterson_Retired — Thu, 21 Apr 2016 01:46:12 GMT

have you ruled out reordering your fields, or for such a small file, producing a new one, with the fields in hierachical order to begin with. That would sure speed up (in terms of code writing and testing) and simplify the concatenation issue.

Re: Alternative for Wildcards in Update Cursor

DonDailey — Thu, 21 Apr 2016 13:17:53 GMT

My first thought was, like you mention, to figure out a way to order the route field so that the routes would be concatenated in the right order to begin with. But the method I used to get the attributes was to create a point class using the intersect tool on our route lines and then doing a spatial join of the points and the route lines they touch to get the route names. I couldn't see a way of forcing the route names to be joined in a particular order. I'm assuming that even if the field values are grabbed in the same direction for each spatial join (clockwise or north to south, etc.) the result would be basically a random ordering because the intersecting routes come from all different directions from intersection to intersection.

Re: Alternative for Wildcards in Update Cursor

DonDailey — Thu, 21 Apr 2016 14:14:55 GMT

It looks like this is going to work. Thanks for the help Darren.

Re: Alternative for Wildcards in Update Cursor

DonDailey — Thu, 21 Apr 2016 14:16:05 GMT

Thanks for that Dan. I originally didn't see any way to do the code correctly.

Re: Alternative for Wildcards in Update Cursor

DarrenWiens2 — Thu, 21 Apr 2016 18:39:58 GMT

Just small edits:

- should be:

for rd in cleaned:

- and

elif rdup[0] == "U": key = 1 # US route

Re: Alternative for Wildcards in Update Cursor

curtvprice — Thu, 21 Apr 2016 23:00:03 GMT

Thanks Darren. Very much appreciate that - hate to have broken code posted!

Re: Alternative for Wildcards in Update Cursor

DonDailey — Sun, 12 Dec 2021 06:29:06 GMT

This is what I ended up using, a combo of the two code suggestions. It slotted the routes into the fields in this hierarchy: Interstate > US > State > Local. Thanks for the help. My python scripting knowledge isn't deep and I learned some things in figuring this one out.

fieldList = ["AllRoutes","ID_Road1_Cross","ID_Road2_Cross","ID_Road3_Cross","ID_Road4_Cross","ID_Road5_Cross"]
   
    with arcpy.da.UpdateCursor(fc,fieldList) as cursor:
        for row in cursor:
              cleaned = [x.strip() for x in row[0].split("&") ]
              rds = []
              for rd in cleaned:
                rdup = rd.upper()
                if rdup.startswith('INTER'): key = 0
                elif rdup.startswith('US'): key = 1 # US route
                elif rdup.startswith('STATE'): key = 2 # state route
                else: key = 3 #local routes
                rds.append([key, rd])
                rds.sort()
              if len(rds) == 2:
                row[1] = rds[0][1]
                row[2] = rds[1][1]
              elif len(rds) == 3:
                row[1] = rds[0][1]
                row[2] = rds[1][1]
                row[3] = rds[2][1]
              elif len(rds) == 4:
                row[1] = rds[0][1]
                row[2] = rds[1][1]
                row[3] = rds[2][1]
                row[4] = rds[3][1]
              elif len(rds) == 5:
                row[1] = rds[0][1]
                row[2] = rds[1][1]
                row[3] = rds[2][1]
                row[4] = rds[3][1]
                row[5] = rds[4][1]
              cursor.updateRow(row)