Skip to content

Make datafusion read parquet folders if non parquet files exists #16460

Closed
@comphead

Description

@comphead

Create test data

ls -la /tmp/t1

-rw-r--r--@  1 xxx  wheel   12 Jun  6 08:35 .part-00000-e248d995-5eac-404e-a2ed-0eb16e27c005-c000.snappy.parquet.crc
-rw-r--r--@  1 xxx  wheel  455 Jun  6 08:35 part-00000-e248d995-5eac-404e-a2ed-0eb16e27c005-c000.snappy.parquet
  • should realize this is a folder and rewrite as /tmp/t1/*.parquet
DataFusion CLI v47.0.0
CREATE EXTERNAL TABLE t1 STORED AS PARQUET location '/tmp/t1';
Parquet error: Parquet error: Invalid Parquet file. Corrupt footer

Originally posted by @comphead in #13456

Metadata

Metadata

Assignees

Labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions