Hello folks,
I am performing some tests with Hive 0.12.0 (cdh5.0.3). I have a quite
complex data model, in particular I modeled a filed in my table as an array
of structs, like:
people array<
struct<
name:string,
surname:string,
address:string,
role:string,
dateofbirth: int,
id: string>
>
I am able to query such field by using the general default UDF function
"array_contains(Array <T>, value)", by hitting something like: select *
from table1 where array_contains(people.name, "Roberto");
What I have experienced is that such function performs a 1:1
match/comparison, and this is fine for some problems. But, (how) can I use
a regular expression applied the sub-field name of my struct (inside my
array people) in order to retrieve e.g. all the people whose name starts
with "Ro"? I know that Hive gives us the "like" and "rlike" functions, but
how can I apply them to a field inside a struct that is one of the elements
of an array?
Please, do not just tell me to change my data model: I'm already
considering this, but the problem is that my table is way more complex (it
is made of several more fields, arrays and arrays of structs).
Thank you.
Roberto
