File path options
For GCS and Amazon S3, users can specify either a single file name as input or provide a folder containing multiple files as input.
If sending a single file, then specify the full path of a json/csv/parquet file inside the bucket.
path=’<folder1>/file.csv’
path=’<folder1>/file.parquet
path=’file.csv’
- Files cannot be compressed
- Files have an extension of csv, json or parquet.
- Csv files should have the header line.
If a folder contains multiple files that are to be used as input, then specify the path of the folder inside the bucket, and ensure
- Path does not have a trailing slash
- All files in the folder have the same extension, either csv/json/parquet.
- All csv files should have the same header line
If folder2 contains all input files, then
path=”<folder1>/<folder2>”
Last modified 1yr ago