updateCursor row[n+1] from searchCursor row[n]-row[n+1]

[JamesCrandall]

FileGDB
ArcGIS 10.4

Need to calculate elapsedSeconds column by first sorting on location_timestamp and then evaluate if PEP_land_name matches in row[n] and row[n+1]. If true, then set row[n+1] to elapsedSeconds value.

fields = ['PEP_land_name','location_timestamp','elapsedSeconds']
with arcpy.da.UpdateCursor(tracksToProcMonth, fields, sql, sql_clause=(None, 'ORDER BY location_timesatmp ASC')) as u_cur:
	with arcpy.da.SearchCursor(tracksToProcMonth,fields, sql, sql_clause=(None, 'ORDER BY location_timestamp ASC')) as cursor:
	  for row in cursor:
		firstlLand = row[0]
		firstTimestamp = row[1]
		try:
		  nextLand = cursor.next()[0]
		  nextTimestamp = cursor.next()[1]

		  if firstlLand == nextLand:
			tdt = nextTimestamp - firstTimestamp
			totSecs = tdt.total_seconds()
			print 'totSecs: {}'.format(totSecs)
			u_row = totSecs
			u_cur.updateRow()

		  else:
			totSecs = 0.00
			u_row = totSecs
			u_cur.updateRow()

I seem to be close in that I can arrive at the correct sort and evaluation, but I'm not exactly sure what to do about the updateCursor and setting the elapsedSeconds value in row[n+1].  In the attached image, OID 1 elapsedSeconds would be 0, OID 5 would be 5, and so on and on...

[Anonymous User]

u_cur.updateRow(row)  Might be what you are after?

... You can move it out from the if/else conditionals because you are setting u_row either way.

        if firstlLand == nextLand:
            tdt = nextTimestamp - firstTimestamp
            totSecs = tdt.total_seconds()
            print 'totSecs: {}'.format(totSecs)
            u_row = totSecs

        else:
            totSecs = 0.00
            u_row = totSecs

        u_cur.updateRow(u_row)

[JamesCrandall]

if the rows oid=1 and oid=5 (see image in op) are the 2 rows being evaluated and elapsedSeconds calculated, does your suggestion update row OID=5?

In the image OID=1 should be 0.00 seconds and OID=5 should be 5.00 seconds

Thanks!

[Anonymous User]

The first thing I saw from your code sample was u_row missing from the .updateRow().  I can't really answer your question without running it through the debugger and verifying that the rest of your code is valid. I can say that it will update whichever row 'u_row' is currently on.

[JamesCrandall]

right... I think that's my problem: I cannot update row[n+1] rather than the current row.

[Anonymous User]

You could probably work the values into a dictionary and do the processing outside of the cursors and then come back to update the values by OID.  I can work up a sample latter today if that is feasible.

[JamesCrandall]

I'm also incorrectly implementing cursor.next()

It seems each time that is invoked it immediately skips to the next row.  I'm not exactly sure how best to get the values for row n and n+1 WHILE the current row is at n.

[DavidPike · Accepted solution]

Unless I've misinterpreted what you're trying to achieve, I think it can be done without and .next() or nested cursors.

I've just created global variables to hold the previous land and timestamp values for each row:

import arcpy


fc = r'PathtoMyFC'
previous_timestamp = ''
previous_land = ''

with arcpy.da.UpdateCursor(fc, ['landname', 'timestamp', 'elapsed']) as cursor:
    for row in cursor:
   
        this_timestamp = row[1]
        current_land = row[0]

        if previous_timestamp is not None and current_land == previous_land:
            time_diff = this_timestamp - previous_timestamp
            total_seconds = time_diff.total_seconds()
            row[2] = total_seconds

        previous_timestamp = this_timestamp
        previous_land = current_land
        cursor.updateRow(row)

[JamesCrandall]

very promising.

...and simple.  Love that.

Let me give it a go and check back but def looks preferable to nested cursors.

Thanks,

j

[JamesCrandall]

I will go ahead and mark as the solution because it appears to work as intended and I am running into another issue that needs resolved in order to actually implement this.

Thanks again!