extracting a tuple from a bag - hadoop

Question

I have a relation of bags of tuples which looks like this. The tuples in the bag come preordered.
{(123,1383313457523,1,US),(123,1383313457543,2,US),(123,1383313457553,3,US)}
{(456,1383313457623,1,UK),(456,1383313457643,2,UK),(456,1383313457653,3,UK)}
{(789,1383313457723,1,UK),(789,1383313457743,2,UK),(789,1383313457753,3,UK)}
Where the tuple is: (id:chararray,time:long,event:chararray,location,chararray)
I want to get the first element of each bag. So my expected output would be:
(123,1383313457523,1,US)
(456,1383313457623,1,UK)
(789,1383313457723,1,UK)
I tried this:
data = load 'mydata.txt' USING PigStorage('\t');
A = FOREACH data GENERATE $0;
dump A;
Which produces the same list of data bags as I had originally.
Alternatively trying to extract just the ids
data = load 'mydata.txt' USING PigStorage('\t');
A = FOREACH data GENERATE $0.$0;
dump A;
I expect:
(123)
(456)
(789)
but I get
{(123),(123),(123)}
{(456),(456),(456)}
{(789),(789),(789)}
How do I adjust my script to get the data that I want.

Answer 1

Use LIMIT inside a nested foreach:
A = FOREACH data { first = LIMIT $0 1; GENERATE FLATTEN(first); }
You cannot count on the tuples in your bag being ordered, since by definition a bag is unordered. However, you can also put an ORDER BY in a nested foreach:
A = FOREACH data { ord = ORDER $0 BY $1; first = LIMIT ord 1; GENERATE FLATTEN(first); }
I find these to be more readable if they are split up onto multiple lines:
A =
FOREACH data {
ord = ORDER $0 BY $1;
first = LIMIT ord 1;
GENERATE
FLATTEN(first);
};
I'm assuming that the bag is ordered by the second field of each tuple ($1).