CODEWITHSUNDEEP
pysparkapache-spark-sql
TJE
TJE

Reputation: 580

How to change values in a PySpark dataframe based on a condition of that same column?

Consider an example dataframe:

df = 
+-------+-----+
|   tech|state|
+-------+-----+
|     70|wa   |
|     50|mn   |
|     20|fl   |
|     50|mo   |
|     10|ar   |
|     90|wi   |
|     30|al   |
|     50|ca   |
+-------+-----+

I want to change the 'tech' column such that any value of 50 is changed to 1 and all other values are equal to 0.

The output would look like this:

df = 
+-------+-----+
|   tech|state|
+-------+-----+
|     0 |wa   |
|     1 |mn   |
|     0 |fl   |
|     1 |mo   |
|     0 |ar   |
|     0 |wi   |
|     0 |al   |
|     1 |ca   |
+-------+-----+

Here's what I have so far:

from pyspark.sql.functions import UserDefinedFunction
from pyspark.sql.types import StringType


changing_column = 'tech'
udf_first = UserDefinedFunction(lambda x: 1, IntegerType())
udf_second = UserDefinedFunction(lambda x: 0, IntegerType())
first_df = zero_df.select(*[udf_first(changing_column) if column == 50 else column for column in zero_df])
second_df = first_df.select(*[udf_second(changing_column) if column != 50 else column for column in first_df])
second_df.show()

Upvotes: 0

Views: 502

Answers (1)

thePurplePython
thePurplePython

Reputation: 2767

hope this helps

from pyspark.sql.functions import when

df = spark\
.createDataFrame([\
    (70, 'wa'),\
    (50, 'mn'),\
    (20, 'fl')],\
    ["tech", "state"])

df\
.select("*", when(df.tech == 50, 1)\
        .otherwise(0)\
        .alias("tech"))\
.show()

+----+-----+----+
|tech|state|tech|
+----+-----+----+
|  70|   wa|   0|
|  50|   mn|   1|
|  20|   fl|   0|
+----+-----+----+

Upvotes: 1

Related Questions