If I define *.tsv files on Amazon S3 as a source for an Athena table and use OpenCSVSerde or LazySimpleSerDe as a deserializer it works correctly. But if I define *.tar.gz files that include *.tsv files I see several strange rows in a table (e.g. a row that contains tsv file name and several empty rows). What is the right way to use tar.gz files in Athena?
1 Answers
The problem is tar, it adds additional rows. Athena can open only *.gz files, but not tar. So in this case I have to use *.gz instead of *.tar.gz.
Related