ArcGIS Functions

gistfile1.txt

# Define a helper function to fix truncated zeros, with one parameter: the value to be processed
def fix_trunc_zeros(val):
    # Use an if statement to check if there are four characters in the string representation of the value
    if len(str(val)) == 4:
        # If this is the case, return the value with an appended "0" in the front
        return "0"+str(val)
    # Otherwise...
    else:
        # Return the value itself
        return str(val)
        
        
 the built-in groupby function for the FIPS and year fields, which you use to group the data by candidate
# Use unstack to perform the table pivot, which will rotate the table and turn rows into columns
df_out = data_df.set_index(['FIPS', 
                            'year', 
                            'county', 
                            'state', 
                            'state_po', 
                            'office', 
                            data_df.groupby(['FIPS', 'year']).cumcount()+1]).unstack()

# Use the indexes for the columns to set column names (Ex: candidate_1, candidate_2, votes_1, votes_2, etc.)
df_out.columns = df_out.columns.map('{0[0]}_{0[1]}'.format)

# Rename columns 
df_out = df_out.rename(columns={"candidate_1": "candidate_dem",
                                "candidatevotes_1": "votes_dem",
                                "candidate_2": "candidate_gop",
                                "candidatevotes_2": "votes_gop",
                                "totalvotes_1": "votes_total",
                                "state_po": "state_abbrev"
                                })

# Keep only the necessary columns
df_out = df_out[["candidate_dem", "votes_dem",
                 "candidate_gop", "votes_gop",
                 "votes_total"]]

# Remove the multiindex since we no longer need these fields to be "locked" for the pivot
df_out.reset_index(inplace=True)

# Print out the first few records to confirm everything worked
df_out.head()