Franco Piccolo
Franco Piccolo

Reputation: 7410

Failed to create table: Error while reading data, error message: Input file is not in Parquet format

Trying to load a *.parquet.gz file as Parquet in BigQuery I get this issue.

Isn't Bigquery supposed to recognize that this is a compressed parquet file?

When I decompress it and load it as .parquet it works.

Upvotes: 2

Views: 1230

Answers (1)

Ricco D
Ricco D

Reputation: 7287

Compressing a parquet with gzip defeats most of the benefits of Parquet columnar compression and reduces our ability to process or parallelize the import.

What BigQuery supports is compression of data blocks on parquet not the whole file itself.

Upvotes: 3

Related Questions