I have sample data as in below string format in Hive table

+———————-+

| col1 |

+———————-+

| 160-80-40 sec|

| 160-80-40 sec|

| 10-10-10-20-20-30 min|

| 10-10-10-20-20-30 min|

| 10-20-30-40-50-60 min|

| 200-100-100 sec|

| 400 200|

+———————-+

I need to find the sum of array, max value, min value

when I load this hive table in spark, it will infer the datatype to string

val df = spark.sql("select col1 from table1")

I tried to use split and size in spark dataframe,
first I split int and string using split function
then again used split with “-” and calculate the sum of array

as it is string type, got error
org.apache.spark.sql.AnalysisException: cannot resolve ‘sum(split(`new_col1, ‘-‘))’ due to data type mismatch: function sum requires numeric types, not array;;

val df1 = df.withColumn("new_col1", split('col1, " ")(0))

df1.withColumn("new_col2", sum(split('new_col1, "-")))

is there a way to cast array[string] to array[int] in dataframe
I would like to solve this problem without using UDFs