I started to pull GLUE table via using pyathena since last week. However, one annoying thing I noticed that is if I wrote my code as shown below, sometimes it works and returns a pandas dataframe but other times, this piece of codes will create a csv and a csv metadata in the folder where physical data (parquet) are stored in S3 and registered in GLUE.
I know that if you use pandas cursor, it may end up with these two files but I just wonder if I can access data without these two files since every time these two files generated in S3, my read in process failed.
Thank you!
import os
access_key_id = os.getenv('AWS_ACCESS_KEY_ID')
secret_access_key = os.getenv('AWS_SECRET_ACCESS_KEY')
connect1 = connect(s3_staging_dir='s3://xxxxxxxxxxxxx')
df = pd.read_sql("select * from abc.table_name", connect1)
df.head()