I have some Hadoop SequenceFiles that contain key/value pairs of 2 Java
classes. The 2 classes each have many Java primitive types.

 

Class A - this is the key

Class B - this is the value

 

I need to store this in a table.

 

As I understand it, I have to:

1)      Store this in an external table

2)      Write a Deserializer for A and B

 

I looked through the "test" examples for various deserializers, but they
usually used Thrift or binary, which does not seem correct for this
purpose.

 

a)      Is there an example available that shows the minimum work needed
to be done?

b)      Does the example show the syntax of the CREATE for the external
table using a SequenceFIle?

c)       Is the key ignored as some Hive documentation shows (I need
values from the key)?

d)      Pig allows a LoadFunc UDF to have access to the key/value. Is
there anything as simple in Hive?

 

John Rodriguez

Reply via email to