Athena gzip compression query result has hybrid compressed-decompressed

Question

I'm setting AWS Athena with s3 bucket which has gzipped csv files.

And then query like this

SELECT * FROM "sample_db"."sample_table2" limit 100;

results is different take 1 and 2.

it seems like to mix compression / decompression results.

Is there any way getting result only decompressed result on Athena?

file contents is below:

"title","user_info.client_user_id","user_info.player_id"
"test : csv take 4",,
"title","user_info.client_user_id","user_info.player_id"
"test : csv take 4",,
"title","user_info.client_user_id","user_info.player_id"
"test : csv take 4",,
"title","user_info.client_user_id","user_info.player_id"
"test : csv take 4",,

s3 has only one file test-sample.gz

Query Take 1

Query Take 2

Answer 1

Cause is wrong format query, partitioning for csv and corrupted data.

It is working on directly s3 gz upload in directories.