Hash Partitioning on fields with different data lengths

[arib0nd]

Quick question on the suitability of hash partitioning (or a suitable alternative/work around) for the following scenario -

Assuming I am reading two different data sets which have the same key (lets say employee id).
Data Set A
employee# - char 5
example values "A100 " , " A101"

Data Set B
Employee# - char 8
example values " A100" , " A101 "

If I partition both data sets using Hash partitioning and sort them, will the join later downstream be successful?
By the time I do the join, I have converted both to the same data type.. in this case varchar(4), but at the time of partitioning, since I am trying to do it up front, they were both different (char 5 and char 8 ) .
Would values such as "A100 " and " A100" across both data sets end up in the same partition or not?

Also would I need to wait to do the sort till after I have combined the data to the same data type as well , or can I also sort them up front?

Thanks!

[ray.wurlod]

There's no guarantee that "A100" and " A100" will be on the same partition, since they're different values. So you will need to make them the same, possibly via string_trim() in a Modify stage or Trim() in a Transformer stage.

Values that are identical will end up on the same partition if the same partitioning algorithm (based on the same key(s)) is used.