Spark DataFrames when udf functions do not accept

问题:

I am preparing a DataFrame with an id and a vector of my features to be used later for doing predictions. I do a groupBy on my dataframe, and in my groupBy I am merging couple of columns as lists into a new column:

def mergeFunction(...) // with 14 input variables

val myudffunction( mergeFunction ) // Spark doesn't support this

df.groupBy("id").agg(
   collect_list(df(...)) as ...
   ... // too many of these (something like 14 of them)
).withColumn("features_labels",
  myudffunction(
     col(...)
     , col(...) )
.select("id", "feature_labels")

This is how I am creating my feature vectors and their labels. It has been working for me so far but this is the first time that my feature vector with this method is getting bigger than number 10 which is what at maximum a udf function in Spark accepts.

I am not sure how else I can fix this? Is the size of udf inputs in Spark going to get bigger, am have I understood them incorrectly, or there is a better way?

回答1:

User defined functions are defined for up to 22 parameters. Only udf helper is define for at most 10 arguments. To handle functions with larger number of parameters you can use org.apache.spark.sql.UDFRegistration.

For example

val dummy = ((
  x0: Int, x1: Int, x2: Int, x3: Int, x4: Int, x5: Int, x6: Int, x7: Int, 
  x8: Int, x9: Int, x10: Int, x11: Int, x12: Int, x13: Int, x14: Int, 
  x15: Int, x16: Int, x17: Int, x18: Int, x19: Int, x20: Int, x21: Int) => 1)

van be registered:

import org.apache.spark.sql.expressions.UserDefinedFunction

val dummyUdf: UserDefinedFunction = spark.udf.register("dummy", dummy)

and use directly

val df = spark.range(1)
val exprs =  (0 to 21).map(_ => lit(1))

df.select(dummyUdf(exprs: _*))

or by name via callUdf

import org.apache.spark.sql.functions.callUDF

df.select(
  callUDF("dummy", exprs:  _*).alias("dummy")
)

or SQL expression:

df.selectExpr(s"""dummy(${Seq.fill(22)(1).mkString(",")})""")

You can also create an UserDefinedFunction object:

import org.apache.spark.sql.expressions.UserDefinedFunction

Seq(1).toDF.select(UserDefinedFunction(dummy, IntegerType, None)(exprs: _*))

In practice having a function with 22 arguments is not very useful and unless you want to use Scala reflection to generate these there are maintenance nightmare.

I would either consider using collections (array, map) or struct as an input or divide this into multiple modules. For example:

val aLongArray = array((0 to 256).map(_ => lit(1)): _*)

val udfWitharray = udf((xs: Seq[Int]) => 1)

Seq(1).toDF.select(udfWitharray(aLongArray).alias("dummy"))

回答2:

Just to expand on zero's answer, it is possible to get .withColumn() function to work with a UDF that has more than 10 parameters. Just need to spark.udf.register() the function and then use an expr for the argument for adding the column (instead of a udf).

For example, something like this should work:

def mergeFunction(...) // with 14 input variables

spark.udf.register("mergeFunction", mergeFunction) // make available in expressions

df.groupBy("id").agg(
   collect_list(df(...)) as ...
   ... // too many of these (something like 14 of them)
).withColumn("features_labels",
  expr("mergeFunction(col1, col2, col3, col4, ...)") ) //pass in the 14 column names
.select("id", "feature_labels")

The underlying expression parser seems to handle more than 10 parameters so I don't think you have to resort to passing around arrays to call the function. Also, if they parameters happen to be different data types, arrays would not work very well.