无法在 DBFS 中保存文件

Question

I have took the azure datasets that are available for practice. I got the 10 days data from that dataset and now I want to save this data into DBFS in csv format. I have facing an error:

" No such file or directory: '/dbfs/temp/hive/mytest.csv'"

but on the other hand if I am able to access the path directly from DBFS. This path is correct.

My code:

from azureml.opendatasets import NoaaIsdWeather
from datetime import datetime
from dateutil import parser 
from dateutil.relativedelta import relativedelta


spark.sql('DROP Table if exists mytest')
dbutils.fs.rm("dbfs:/tmp/hive",recurse = True)

basepath = "dbfs:/tmp/hive" 

try:
  dbutils.fs.ls(basepath)
except:
  dbutils.fs.mkdirs(basepath)
else:
  raise Exception("The Folder "+ basepath + " already exist, this notebook will remove in the end")

dbutils.fs.mkdirs("dbfs:/tmp/hive")

start_date = parser.parse('2020-5-1')
end_date = parser.parse('2020-5-10')

isd = NoaaIsdWeather(start_date, end_date)
pdf = isd.to_spark_dataframe().toPandas().to_csv("/dbfs/temp/hive/mytest.csv")

What should I do?

Thanks

Answer 1

I tried reproducing the same issue. First I have used the following code and made sure that the directory exists using os.listdir() .

from azureml.opendatasets import NoaaIsdWeather
from datetime import datetime
from dateutil import parser 
from dateutil.relativedelta import relativedelta
spark.sql('DROP Table if exists mytest')
dbutils.fs.rm("dbfs:/tmp/hive",recurse = True)
basepath = "dbfs:/tmp/hive" 
try:
  dbutils.fs.ls(basepath)
except:
  dbutils.fs.mkdirs(basepath)
else:
  raise Exception("The Folder "+ basepath + " already exist, this notebook will remove in the end")

dbutils.fs.mkdirs("dbfs:/tmp/hive")

import os  
os.listdir("/dbfs/tmp/hive/")

• Then I used the following to write the csv using to_pandas_dataframe() . This has successfully written the required dataframe to csv file in required path.

mydf = isd.to_pandas_dataframe()  
mydf.to_csv("/dbfs/tmp/hive/mytest.csv")