fixed dataframe merge when specifying ancillary fields

Author: jpswinski

sliderule/icesat2.py

@@ -842,7 +842,8 @@ def atl06p(parm, asset=DEFAULT_ASSET, version=DEFAULT_ICESAT2_SDP_VERSION, callb
             tstart_merge = time.perf_counter()
             for field in field_dictionary:
                 df = geopandas.pd.DataFrame(field_dictionary[field])
-                gdf = gdf.merge(df, on='extent_id', how='inner').set_axis(gdf.index)
+                gdf = gdf.merge(df, on='extent_id', how='inner')
+                gdf.set_axis(gdf.index)
             profiles["merge"] = time.perf_counter() - tstart_merge
 
             # Delete Extent ID Column

sliderule/sliderule.py

@@ -66,6 +66,8 @@
 
 arrow_file_table = {}
 
+profiles = {}
+
 gps_epoch = datetime(1980, 1, 6)
 tai_epoch = datetime(1970, 1, 1, 0, 0, 10)
 
@@ -263,8 +265,14 @@ def __parse_native(data, callbacks):
     rec_index = 0
     rec_rsps = []
 
+    duration = 0.0
+
     for line in data.iter_content(None):
 
+        # Capture Start Time (for duration)
+        tstart = time.perf_counter()
+
+        # Process Line Read
         i = 0
         while i < len(line):
 
@@ -315,6 +323,12 @@ def __parse_native(data, callbacks):
                 rec_size_index = 0
                 rec_index = 0
 
+        # Capture Duration
+        duration = duration + (time.perf_counter() - tstart)
+
+    # Update Timing Profile
+    profiles[__parse_native.__name__] = duration
+
     return recs
 
 #

utils/utils.py

@@ -1,5 +1,6 @@
 import time
 import json
+import sliderule
 from sliderule import icesat2
 
 #
@@ -155,6 +156,20 @@ def initialize_client(args):
     # Return Parameters and Configuration
     return parms, cfg
 
+#
+# Display Timing
+#
+def display_timing():
+
+    print("\nSlideRule Timing Profiles")
+    for key in sliderule.profiles:
+        print("{:20} {:.6f} secs".format(key + ":", sliderule.profiles[key]))
+
+    print("\nICESat2 Timing Profiles")
+    for key in icesat2.profiles:
+        print("{:20} {:.6f} secs".format(key + ":", icesat2.profiles[key]))
+
+
 #
 # Display Statistics
 #
@@ -171,9 +186,7 @@ def display_statistics(gdf, name):
     else:
         print("No {} were returned".format(name))
 
-    print("\nTiming Profiles")
-    for key in icesat2.profiles:
-        print("{:16}: {:.6f} secs".format(key, icesat2.profiles[key]))
+    display_timing()
 
 #
 # Pretty Print JSON