Calculate Percentile Ranks by Group using Numpy - numpy

Question

I'm very new with Python, and I want to calculate percentile ranks by group. My group is wildlife management unit (WMU - string), and ranks are based the value of predicted moose density (PMDEN3 - FLOAT). The rank value goes into the field RankMD.
My approach was to use the for loop to calculate the 3 ranks within each WMU, but the result is that 3 ranks are created for the entire dbf file (about 23,000 records), without respect to WMU. Any help is much appreciated.
import arcpy
import numpy as np
input = r'K:\Moose\KrigStratPython\TestRank3.dbf'
arr = arcpy.da.TableToNumPyArray(input, ('PMDEN3', 'Wmu'))
c_arr = [float(x[0]) for x in np.ndarray.flatten(arr)]
for Wmu in arr:
##to create 3 rank for example
p1 = np.percentile(c_arr, 33) # rank = 0
p2 = np.percentile(c_arr, 67) # rank = 1
p3 = np.percentile(c_arr, 100) # rank = 2
#use cursor to update the new rank field
with arcpy.da.UpdateCursor(input , ['PMDEN3','RankMD']) as cursor:
for row in cursor:
if row[0] < p1:
row[1] = 0 #rank 0
elif p1 <= row[0] and row[0] < p2:
row[1] = 1
else:
row[1] = 2
cursor.updateRow(row)

Answer 1

Your for loop is correct, however, your UpdateCursor is iterating over all rows in the table. To get your desired result you need to select out a subset of the table, and then use the update cursor on that. You can do this by passing a query to the where_clause parameter of the UpdateCursor function.
So you would have a query like this:
current_wmu = WMU['wmu'] # This should be the value of the wmu that the for loop is currently on I think it would be WMU['wmu'] but i'm not positive
where_clause = "WMU = '{}'".format(current_wmu) # format the above variable into a query string
and then your UpdateCursor would now be:
with arcpy.da.UpdateCursor(input , ['PMDEN3','RankMD'], where_clause) as cursor:

Answer 2

Based on suggestion from BigGerman, I revised my code and this is now working. Script loops through each WMU value, and calculates rank percentile within each group based on PMDEN. To improve the script I should create an array of WMU values from my input file rather than manually creating the array.
import arcpy
import numpy as np
#fields to be calculated
fldPMDEN = "PMDEN"
fldRankWMU = "RankWMU"
input = r'K:\Moose\KrigStratPython\TestRank3.dbf'
arcpy.MakeFeatureLayer_management(input, "stratLayerShpNoNullsLyr")
WMUs = ["10", "11A", "11B", "11Q", "12A"]
for current_wmu in WMUs:
##to create 3 rank for example
where_clause = "Wmu = '{}'".format(current_wmu) # format the above variable into a query
with arcpy.da.UpdateCursor("stratLayerShpNoNullsLyr", [fldPMDEN,fldRankWMU], where_clause) as cursor:
arr1 = arcpy.da.TableToNumPyArray("stratLayerShpNoNullsLyr", [fldPMDEN,fldRankWMU], where_clause)
c_arrS = [float(x[0]) for x in np.ndarray.flatten(arr1)]
p1 = np.percentile(c_arrS, 33) # rank = 3
p2 = np.percentile(c_arrS, 67) # rank = 2
p3 = np.percentile(c_arrS, 100) # rank = 1 (highest density)
for row in cursor:
if row[0] < p1:
row[1] = 3 #rank 0
elif p1 <= row[0] and row[0] < p2:
row[1] = 2
else:
row[1] = 1
cursor.updateRow(row)