猪代数与弦

Question

I need to write an interface that would collapse individual items in a bag to a string. Example: {(a),(b)} into ab.

First, can algebraic interface be used to return strings or is it restricted to integers (Long).

Second, is there a place where I can access some examples of using algebraic (apart from the COUNT example I see everywhere).

Just answer if Algebraic can be used to process strings and let me know if there is a good place where I can see some existing UDF code (not the ones that exactly solve my problem).

These are the things I have tried:

1. Googling for any UDF code that works on bags. Not getting anything apart from the COUNT example that is posted everywhere
2. Trying out different options in Pig. Apparently you cannot de-reference individual items inside a bag which is a bummer.

Answer 1

Finally this is what I figured out:

1. If your problem can be solved by JOIN as efficiently as GROUP, you should do it. GROUPs create bags which are harder to deal with.

2. You are not obligated to use ALGEBRAIC to deal with bags. Instead you can just write a EVAL UDF. However it is going to be much slower if your bag size is large.