Garrett Weaver created ARROW-14959:
--------------------------------------
Summary: Reading Hive-style partitioned parquet files from GCS
Key: ARROW-14959
URL: https://issues.apache.org/jira/browse/ARROW-14959
Project: Apache Arrow
Issue Type: Bug
Affects Versions: 6.0.1
Reporter: Garrett Weaver
Trying to read a spark-generated hive-style partitioned parquet dataset with
`gcsfs` and `pyarrow`, but getting a `FileNotFoundError` if I try to read from
the base directory or even if try to read directly from one of the partitions.
Not sure if I am doing something wrong or it is not supported.
Note that I have successfully read this hive-style partitioned parquet dataset
using other methods to rule out any other issues, including:
* Successful read with pyspark by using spark.read.parquet
* Successful read of a specific partition by passing list of paths to
`ParquetDataset`
* Also, tested reading another spark-generated parquet dataset with no
Hive-style partitions from GCS and that worked as well.
Below is what I am trying:
```
import gcsfs
import pyarrow as pa
gcs = gcsfs.GCSFileSystem()
partitions = pa.dataset.partitioning(
pa.schema([("partition_var", pa.string())]), dictionaries="infer",
flavor="hive"
)
partitioned_dataset = pa.dataset.dataset(
"path/to/partitioned/dataset/base/dir",
filesystem=gcs,
format="parquet",
partitioning=partitions,
)
partition_of_dataset = pa.dataset.dataset(
"path/to/partitioned/dataset/base/dir/partition_var=some_value",
filesystem=gcs,
format="parquet",
)
```
The errors returned for both are below:
```
FileNotFoundError: path/to/partitioned/dataset/base/dir/
FileNotFoundError:
path/to/partitioned/dataset/base/dir/partition_var=some_value/
```
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
