Hi, you can inspect the avro files with avro-tools utility.
create table work.test_avro ( i int, s string ) stored as avro;
insert into work.test_avro select 1, "abc";
set hive.exec.compress.output = true
set hive.exec.compress.intermediate = true;
set avro.output.codec= snappy;
insert into work.test_avro select 2, "abcdefgb";
In this table there are two file, one compressed with snappy one without compression, you can check it with get-meta command:
$ avro-tools getmeta 000000_0
avro.schema {"type":"record","name":"test_avro","namespace":"work","fields":[{"name":"i","type":["null","int"],"default":null},{"name":"s","type":["null","string"],"default":null}]}
$ avro-tools getmeta 000000_0_copy_1
avro.schema {"type":"record","name":"test_avro","namespace":"work","fields":[{"name":"i","type":["null","int"],"default":null},{"name":"s","type":["null","string"],"default":null}]}
avro.codec snappy