Take2

take2.py

from pyspark.sql.functions import udf, col

@udf('string')
def get_zip_udf2(latitude, longitude):
    search = SearchEngine()
    try:
        zip = search.by_coordinates(latitude, longitude, returns=1)[0].to_dict()["zipcode"]
    except:
        zip = 'bad'
    return zip

  
df.withColumn('zip', get_zip_udf2(col("latitude"),col("longitude"))).show()

# +---+---------+-----------+---+
# | id| latitude|  longitude|zip|
# +---+---------+-----------+---+
# |  1|33.704045| -90.754334|bad|
# |  2|45.019704|-123.014528|bad|
# |  3|26.306754| -80.259649|bad|
# +---+---------+-----------+---+