fixed duplicate handling for points

USEPA · Apr 18, 2022 · 5b6a86b · 5b6a86b
1 parent d197e00
commit 5b6a86b
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 6 deletions.
diff --git a/LakeCat.py b/LakeCat.py
@@ -16,7 +16,7 @@
 
 from border import makeBrdrPctFile
 from lake_cat_config import LYR_DIR, NHD_DIR, OUT_DIR, STREAMCAT_DIR, FRAMEWORK
-from LakeCat_functions import (Accumulation, NHDtblMerge, PointInPoly,
+from LakeCat_functions import (Accumulation, NHDtblMerge, PointInPoly,dbf2DF,
                                chkColumnLength, doStats, getOnNetLakes, inputs,
                                makeBasins, makeNParrays, rpus)
 

diff --git a/LakeCat_functions.py b/LakeCat_functions.py
@@ -435,11 +435,10 @@ def PointInPoly(points, inZoneData, pct_full, fld="GRIDCODE", summaryfield=None)
     if not points.crs == polys.crs:
         points = points.to_crs(polys.crs)
     # Get list of lat/long fields in the table
-    latlon = [
-        s for s in points.columns if any(xs in s.upper() for xs in ["LONGIT", "LATIT"])
-    ]
-    # Remove duplicate points for 'Count'
-    points2 = points.loc[~points.duplicated(latlon)]
+    points['lon'] = points['geometry'].x
+    points['lat'] = points['geometry'].y
+    # Remove duplicate points
+    points2 = points.drop_duplicates(subset=["lon", "lat"], keep='last')
     try:
         point_poly_join = sjoin(points2, polys, how="left", op="within")
     except: