Converting latitude and longitude to UTM coordinates in pyspark

Question

I have dataframe contain longitude and latitude coordinates for each point. I want to convert the geographical coordinates for each point to UTM coordinates.

I tried to use utm module (https://pypi.org/project/utm/)

import utm
df=df.withColumn('UTM',utm.from_latlon(fn.col('lat'),fn.col('lon')))

but I obtain this error :

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-39-8b21f98738ca> in <module>()
----> 1 df=df.withColumn('UTM',utm.from_latlon(fn.col('lat'),fn.col('lon')))

~\Anaconda3\lib\site-packages\utm\conversion.py in from_latlon(latitude, longitude, force_zone_number)
    152        .. _[1]: http://www.jaworski.ca/utmzones.htm
    153     """
--> 154     if not -80.0 <= latitude <= 84.0:
    155         raise OutOfRangeError('latitude out of range (must be between 80 deg S and 84 deg N)')
    156     if not -180.0 <= longitude <= 180.0:

F:\spark\spark\python\pyspark\sql\column.py in __nonzero__(self)
    633 
    634     def __nonzero__(self):
--> 635         raise ValueError("Cannot convert column into bool: please use '&' for 'and', '|' for 'or', "
    636                          "'~' for 'not' when building DataFrame boolean expressions.")
    637     __bool__ = __nonzero__

ValueError: Cannot convert column into bool: please use '&' for 'and', '|' for 'or', '~' for 'not' when building DataFrame boolean expressions.

update:

After creating udf that applying utm or pyproj function

The result is:

+--------------------+
|                 UTM|
+--------------------+
|[Ljava.lang.Objec...|
|[Ljava.lang.Objec...|
|[Ljava.lang.Objec...|
|[Ljava.lang.Objec...|
|[Ljava.lang.Objec...|
+--------------------+
only showing top 5 rows

mayank agrawal mayank agrawal · Accepted Answer · 2018-06-28T07:40:32

Something like this,

import pyspark.sql.functions as F
import utm
from pyspark.sql.types import *

utm_udf_x = F.udf(lambda x,y: utm.from_latlon(x,y)[0], FloatType())
utm_udf_y = F.udf(lambda x,y: utm.from_latlon(x,y)[1], FloatType())

df = df.withColumn('UTM_x',utm_udf_x(F.col('lat'), F.col('lon')))
df = df.withColumn('UTM_y',utm_udf_y(F.col('lat'), F.col('lon')))

Although I am not sure why did you write [1] at the end.

Converting latitude and longitude to UTM coordinates in pyspark

1 Answers