聚合函数 Spark SQL

Question

I am working with Pyspark 2.2 rn, my code crashes by this function and I don't understand why it's crashed on ->

code

import Pyspark.sql.functions as F
t.withColumns('column_name',
              F.expr("aggregate(column, '', (acc, x) -> acc || concat(x, 4) ','))"))

error like - "extraneous input '>' expecting 'all of sql functions'"

ty for ur help

Answer 1

Three suggestions.

1. Remove the ',')) : those parentheses are unmatched.

2. Use either …||… or concat(…,…) ; no need for both.

3. Because the column expression contains > and the column is not aliased, spark is trying to turn the expression itself into the column name, but column names cannot contain > , as the OP's error message alludes to.

   Solution: Alias the column.

   (See screenshot. Yes, the screenshot uses spark SQL instead of pyspark, but the OP's issue is with the spark SQL snippet within the pyspark.)