Better wildcard select on more column Operations

[jayckaiser]

PR110 added a select-all wildcard in ModifyColumnsOperation. By using "*" as the column name, you can apply the same column transformation to every column in the dataframe.

There are two additions I'd like to make to this feature:

1. Allow asterisks to be used as a wildcard capture for partial column name matching, not just all columns. This match could be done using the built-in fnmatch library. For example (thanks Tom):

   # given a $sources.X with columns `a_1`, `a_2`, `b_1`, `b_2`, `a_1_b_2`, `b_1_a_2`
   transformations:
   my_test:
   source: $sources.X
   operations:
   - operation: modify_columns
     columns:
       a_*: "{%raw%}{{value|trim}}{%endraw%}" # modifies `a_1`, `a_2`, `a_1_b_2`
       *_1: "{%raw%}{{value|trim}}{%endraw%}" # modifies `a_1`, `b_1`
       *_1_*: "{%raw%}{{value|trim}}{%endraw%}" # modifies `a_1_b_2`, `b_1_a_2`
       *: "{%raw%}{{value|trim}}{%endraw%}" # modifies `a_1`, `a_2`, `b_1`, `b_2`, `a_1_b_2`, `b_1_a_2` (all columns)

2. Extend wildcard matching to more column operations:

   • ModifyColumnsOperation
   • DropColumnsOperation
   • KeepColumnsOperation
   • MapValuesOperation
   • DateFormatOperation
   • SnakeCaseColumnsOperation