Make XY Event Layer in arcpy using a pandas dataframe

Question

I have a Pandas df with Latitude and Longitude columns I am trying to export to a featureclass using ArcPy. I cannot resolve the fiona and GeoPandas issue so I tried to use MakeXYEventLayer_management and FeatureClassToFeatureclass_conversion

import pandas as pd, numpy as np, calendar
from os.path import splitext
from arcpy import MakeXYEventLayer_management, FeatureClassToFeatureClass_conversion
data = 'Sample_Data_2021.xlsx'
df = pd.read_excel(data)
MakeXYEventLayer_management(df, 'LONGITUDE', 'LATITUDE', 'data_xy', spRef)
FeatureClassToFeatureClass_conversion('data_xy', outDir, outFeatureClass)
Traceback (most recent call last):
 File ".../main.py", line 105, in <module>
MakeXYEventLayer_management(df, 'LONGITUDE', 'LATITUDE', 'data_xy', spRef)
 File "...\management.py", line 9601, in MakeXYEventLayer
raise e
 File "...\management.py", line 9598, in MakeXYEventLayer
retval = convertArcObjectToPythonObject(gp.MakeXYEventLayer_management(*gp_fixargs((table, in_x_field, in_y_field, out_layer, spatial_reference, in_z_field), True)))
 File "...\_base.py", line 512, in <lambda>
return lambda *args: val(*gp_fixargs(args, True))
RuntimeError: Object: Error in executing tool

I tried using df.LATITUDE\df['LATITUDE'] in the XY inputs but got the same error. I could convert it to CSV but then I would loose all the dtypes I assigned. Is there a way to directly convert a Pandas df to a featureclass?

• enolan: This may help you - community.esri.com/t5/arcgis-api-for-python-questions/…
• user2856: Why use a pandas df when you can go straight from Excel to feature class? Note a feature class is not a feature dataset. Feature datasets are collections of feature classes
• Son of a Beach: I expect that the reason they're avoiding using ExcelToTable_conversion() is due to the dtypes (as mentioned in the OP). ArcGIS quite often gets these wrong on importing Excel. Does the schema.ini work-around to specify data types work for Excel, or is it just CSV? I can't remember.
• MrKingsley: The 2 reasons for going to Pandas is to force the dtypes because esri doesn't get them right and I was having difficulties doing the table edits and setting the dtypes outside of Pandas. Since Pandas comes with the ArcPy env I figured I could do my work there. Yes, I am aware of the difference between an esri featureclass and esri featuredataset; I just typed the wrong thing.

Answer 1

I do this using NumPyArrayToFeatureClass. You'll need to convert your df to a np array. I use this to get the dtypes right.

import arcpy
import numpy as np
import pandas as pd
import os
# list data types
list_dtypes = [i.str for i in df.dtypes]
# turn object type into 255 character string type
# turn datetimes type into M8[us] type
list_dtypes = ['<U255' if d == '|O' else d for d in list_dtypes]
list_dtypes = ['M8[us]' if d == '<M8[ns]' else d for d in list_dtypes]
# format as list of tuples (field name and adjusted dtype)
tup_dtypes = [(f,t) for f,t in zip(list(df.columns.values),list_dtypes)]
# write to array, recasting the to_records array with .astype()
array = df.to_records(index=False).astype(tup_dtypes)
# save to feature class
outPath = os.path.join(outDir, outFeatureClass)
arcpy.da.NumPyArrayToFeatureClass(array, outPath, ("LONGITUDE", "LATITUDE"))

• MrKingsley: Is it still arcpy.da.NumPyArrayToFeatureClass? I treied arcpy.da, arcpy.managment, arcpy.NumPyArrayToFeatureClass_management and I keep getting the error Cannot find reference.
• Brennan: Did you update the output parameter? I tweaked the example a bit for you. Here's the link. pro.arcgis.com/en/pro-app/latest/arcpy/data-access/…
• MrKingsley: I got an error but also a shapefile that looks to be complete. Traceback (most recent call last): File ".../main.py", line 124, in <module> da.NumPyArrayToFeatureClass(array, outPath, ("LONGITUDE", "LATITUDE")) SystemError: <built-in function NumPyArrayToFeatureClass> returned NULL without setting an error