Page History

...

• arrays: ARRAY<data_type>
• maps: MAP<primitive_type, data_type>

• Wiki Markup
  structs: {{STRUCT <colSTRUCT<col_name : data_type \[COMMENT col_comment], ...>}}

• union: UNIONTYPE<data_type, data_type, ...>

Timestamps

Supports traditional UNIX timestamp with optional nanosecond precision.

...

Timestamps are interpreted to be timezoneless and stored as an offset from the UNIX epoch. Convenience UDFs for conversion to and from timezones are provided (to_utc_timestamp, from_utc_timestamp).
All existing datetime UDFs (month, day, year, hour, etc.) work with the TIMESTAMP data type.

Union types

Union types can at any one point hold exactly one of their specified data types. You can create an instance of this type using the create_union UDF:

CREATE TABLE union_test(foo UNIONTYPE<int, double, array<string>, struct<a:int,b:string>>);
SELECT foo FROM union_test;

{0:1}
{1:2.0}
{2:["three","four"]}
{3:{"a":5,"b":"five"}}
{2:["six","seven"]}
{3:{"a":8,"b":"eight"}}
{0:9}
{1:10.0}

The first part in the deserialized union is the tag which lets us know which part of the union is being used. In this example 0 means the first data_type from the definition which is an int and so on.

To create a union you have to provide this tag to the create_union UDF:

SELECT create_union(0, key), create_union(if(key<100, 0, 1), 2.0, value), create_union(1, "a", struct(2, "b")) FROM src LIMIT 2;

{0:"238"}	{1:"val_238"}	{1:{"col1":2,"col2":"b"}}
{0:"86"}	{0:2.0}	{1:{"col1":2,"col2":"b"}}

Literals

Integral types

Integral literals are assumed to be INT by default, unless the number exceeds the range of INT in which case it is interpreted as a BIGINT, or if one of the following postfixes is present on the number.

...