SnowflakeDocumentation Home

SPRKSCL1119

org.apache.spark.sql.Column.endsWith

Message: org.apache.spark.sql.Column.endsWith has a workaround, see documentation for more info

Category: Warning

Description

This issue appears when the SMA detects a use of the org.apache.spark.sql.Column.endsWith function, which has a workaround.

Scenario

Input

Below is an example of the org.apache.spark.sql.Column.endsWith function, first used with a literal string argument and then with a column object argument.

val df1 = Seq(
  ("Alice", "alice@example.com"),
  ("Bob", "bob@example.org"),
  ("David", "david@example.com")
).toDF("name", "email")
val result1 = df1.filter(col("email").endsWith(".com"))

val df2 = Seq(
  ("Alice", "alice@example.com", ".com"),
  ("Bob", "bob@example.org", ".org"),
  ("David", "david@example.org", ".com")
).toDF("name", "email", "suffix")
val result2 = df2.filter(col("email").endsWith(col("suffix")))

Output

The SMA adds the EWI SPRKSCL1119 to the output code to let you know that this function is not directly supported by Snowpark, but it has a workaround.

val df1 = Seq(
  ("Alice", "alice@example.com"),
  ("Bob", "bob@example.org"),
  ("David", "david@example.com")
).toDF("name", "email")
/*EWI: SPRKSCL1119 => org.apache.spark.sql.Column.endsWith has a workaround, see documentation for more info*/
val result1 = df1.filter(col("email").endsWith(".com"))

val df2 = Seq(
  ("Alice", "alice@example.com", ".com"),
  ("Bob", "bob@example.org", ".org"),
  ("David", "david@example.org", ".com")
).toDF("name", "email", "suffix")
/*EWI: SPRKSCL1119 => org.apache.spark.sql.Column.endsWith has a workaround, see documentation for more info*/
val result2 = df2.filter(col("email").endsWith(col("suffix")))

Recommended fix

As a workaround, you can use the com.snowflake.snowpark.functions.endswith function, where the first argument would be the column whose values will be checked and the second argument the suffix to check against the column values. Please note that if the argument of the Spark's endswith function is a literal string, you should convert it into a column object using the com.snowflake.snowpark.functions.lit function.

val df1 = Seq(
  ("Alice", "alice@example.com"),
  ("Bob", "bob@example.org"),
  ("David", "david@example.com")
).toDF("name", "email")
val result1 = df1.filter(endswith(col("email"), lit(".com")))

val df2 = Seq(
  ("Alice", "alice@example.com", ".com"),
  ("Bob", "bob@example.org", ".org"),
  ("David", "david@example.org", ".com")
).toDF("name", "email", "suffix")
val result2 = df2.filter(endswith(col("email"), col("suffix")))

Additional recommendations

PreviousSPRKSCL1118NextSPRKSCL1120

Last updated