[PIG-4471] ASSERT Bag is not empty and/or is within a specified size range.

• Type: New Feature
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: internal-udfs, parser
• Assignee: Unassigned
• Reporter: Niels Basjes
• Created: 19/Mar/15 15:49
• Updated: 24/Apr/15 14:52

Description

In PIG-3367 the ASSERT keyword was created.
The current implementation allows for checking in each record in the bag if the value of a column is valid (and fail the job if it is not).

We did several experiments and found that an empty bag (0 tuples) always succeeds. We need to ensure that a bag has been loaded correctly.

Proposed enhancements:

1. Allow the ASSERT statement to check if a bag is empty.

   A = LOAD 'data' AS (a0:int,a1:int,a2:int);
   ASSERT A NOT EMPTY, 'The A bag may not be empty';
   

2. Allow the ASSERT statement to check if a bag has more than (or less than) a specific number of tuples.

   A = LOAD 'data' AS (a0:int,a1:int,a2:int);
   ASSERT SIZE A > 100, 'The A bag is not big enough';
   ASSERT SIZE A < 1000, 'The A bag is too big';
   

   • For me this may be an approximating implementation. i.e. if I say it must have at least 5M tuples then it may still return 'is valid' if it has 4.9M tuples.

NOTE: The syntax I show is just to give an idea on what I want to do.