rdblue commented on code in PR #41:
URL: https://github.com/apache/iceberg-python/pull/41#discussion_r1451834512
##########
pyiceberg/io/pyarrow.py:
##########
@@ -1565,13 +1564,54 @@ def fill_parquet_file_metadata(
del upper_bounds[field_id]
del null_value_counts[field_id]
- df.file_format = FileFormat.PARQUET
df.record_count = parquet_metadata.num_rows
- df.file_size_in_bytes = file_size
df.column_sizes = column_sizes
df.value_counts = value_counts
df.null_value_counts = null_value_counts
df.nan_value_counts = nan_value_counts
df.lower_bounds = lower_bounds
df.upper_bounds = upper_bounds
df.split_offsets = split_offsets
+
+
+def write_file(table: Table, tasks: Iterator[WriteTask]) -> Iterator[DataFile]:
+ task = next(tasks)
+
+ try:
+ _ = next(tasks)
+ # If there are more tasks, raise an exception
+ raise ValueError("Only unpartitioned writes are supported:
https://github.com/apache/iceberg-python/issues/208";)
+ except StopIteration:
+ pass
+
+ df = task.df
+
+ file_path =
f'{table.location()}/data/{_generate_datafile_filename("parquet")}'
+ file_schema = schema_to_pyarrow(table.schema())
+
+ collected_metrics: List[pq.FileMetaData] = []
+ fo = table.io.new_output(file_path)
+ with fo.create() as fos:
+ with pq.ParquetWriter(fos, schema=file_schema, version="1.0",
metadata_collector=collected_metrics) as writer:
+ writer.write_table(df)
+
+ df = DataFile(
+ content=DataFileContent.DATA,
+ file_path=file_path,
+ file_format=FileFormat.PARQUET,
+ partition=Record(),
+ record_count=len(df),
+ file_size_in_bytes=len(fo),
+ # Just copy these from the table for now
+ sort_order_id=table.sort_order().order_id,
+ spec_id=table.spec().spec_id,
Review Comment:
Since `write_file` is a public method, we can't guarantee that the caller
did this check. I agree that it is safe when called from `append` or
`overwrite`, but a caller could use this method directly to create a data file
for a partitioned table right?
Wouldn't it be easy to just pass the spec ID and partition tuple (an empty
`Record`) through `WriteTask` for now? I think it would make sense if a
`WriteTask` were for a single partition.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
