python - h2o 并从 python 设置目标框架

Question

我们使用 python 与单实例 h2o 对话（最新版本 3.22.1.1）。

有时我们会收到此错误：

DistributedException from /10.192.21.17:54321: 'class water.fvec.Frame s3a://BUCKET_NAME/part-00001-0cd59acc-d03f-4af6-8227-e58bf7ad9562-c000.snappy.parquet is already in use.  Unable to use it now.  Consider using a different destination name.', caused by java.lang.IllegalArgumentException: class water.fvec.Frame s3a://BUCKET_NAME/part-00001-0cd59acc-d03f-4af6-8227-e58bf7ad9562-c000.snappy.parquet is already in use.  Unable to use it now.  Consider using a different destination name.
    at water.MRTask.getResult(MRTask.java:478)
    at water.MRTask.getResult(MRTask.java:486)
    at water.MRTask.doAll(MRTask.java:402)

我们尝试像这样传递我们的随机destination_frame：

h2o.import_file(
                path=data_path,
                destination_frame='frame_{}'.format(str(uuid.uuid4())))

但看起来 H2O 没有使用 destination_frame 参数，即使我们在日志中看到它存在：

POST /3/Parse, parms: {number_columns=94, source_frames=["s3a://BUCKET_NAME/part-00000-0cd59acc-d03f-4af6-8227-e58bf7ad9562-c000.snappy.parquet"], column_types=["UUID","Numeric","Numeric","Numeric","Numeric","Numeric","Numeric","Enum","Enum","Time","Numeric","Enum","Enum","Time","Time","Numeric","Enum","Enum","Numeric","Enum","Numeric","Numeric","Numeric","Enum","Enum","Enum","Enum","Enum","Numeric","Enum","Enum","Numeric","Enum","Numeric","Numeric","Numeric","Numeric","Numeric","Numeric","Numeric","Numeric","Time","Numeric","Enum","Enum","Time","Numeric","Numeric","Enum","Enum","Enum","Enum","Enum","Numeric","Enum","Numeric","Enum","Numeric","Enum","Numeric","Enum","Numeric","Enum","Numeric","Numeric","Numeric","Numeric","UUID","Time","Numeric","Numeric","Enum","Numeric","Numeric","Numeric","Enum","Numeric","Numeric","Enum","Enum","Numeric","UUID","Numeric","Numeric","Numeric","Numeric","Numeric","Numeric","Numeric","Numeric","Enum","Numeric","Numeric","Numeric"], single_quotes=True, parse_type=PARQUET, destination_frame=frame_19d32a0b-812f-4179-ba83-c3e1afe1d84f, column_names=[
"ALL_COLUMN_NAMES_HERE"], delete_on_done=True, check_header=1, separator=124, blocking=False, chunk_size=77450}

标签： pythonparqueth2o