CODEWITHSUNDEEP
amazon-web-servicesamazon-s3aws-glue
JOHN
JOHN

Reputation: 1511

aws glue `ImportError: cannot import name 'S3ArnParamHandler'`

I developed a pandas etl script locally and works fine.

I prepared a wheel file and uploaded to s3. All packages are installed properly.

However, when the script run, it shows ImportError: cannot import name 'S3ArnParamHandler'

Below is my requirements.txt

awscli==1.18.140
asn1crypto==1.4.0
awswrangler==1.9.3
azure-common==1.1.25
azure-core==1.8.1
azure-storage-blob==12.5.0; python_version >= '3.5.2'
boto3==1.14.63
botocore==1.17.63
certifi==2020.6.20
cffi==1.14.2
chardet==3.0.4
cryptography==2.9.2; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4'
docutils==0.15.2; python_version >= '2.6' and python_version not in '3.0, 3.1, 3.2'
fsspec==0.8.2
idna==2.9; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'
isodate==0.6.0
jmespath==0.10.0; python_version >= '2.6' and python_version not in '3.0, 3.1, 3.2'
msrest==0.6.19
numpy==1.19.2
oauthlib==3.1.0; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'
oscrypto==1.2.1
packaging==20.4; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'
pandas==1.0.0
psycopg2-binary==2.8.6; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'
pyarrow==1.0.1; python_version >= '3.5'
pycparser==2.20; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'
pycryptodomex==3.9.8; python_version >= '2.6' and python_version not in '3.0, 3.1, 3.2, 3.3'
pyjwt==1.7.1
pymysql==0.9.0
pyopenssl==19.1.0
pyparsing==2.4.7; python_version >= '2.6' and python_version not in '3.0, 3.1, 3.2'
python-dateutil==2.8.1
pytz==2020.1
requests-oauthlib==1.3.0
requests==2.23.0
s3fs==0.4.2
s3transfer==0.3.3
six==1.15.0
snowflake-connector-python==2.3.2; python_version >= '3.5'
snowflake-sqlalchemy==1.2.3
sqlalchemy-redshift==0.8.1
sqlalchemy==1.3.13
urllib3==1.25.10
xlrd==1.2.0

Upvotes: 5

Views: 3809

Answers (3)

pyxai
pyxai

Reputation: 31

I just uninstalled boto3, then install it again

Upvotes: 0

Rich G
Rich G

Reputation: 21

Tomasz is correct - it has to do with the (lower) versions of boto3 / botocore that are currently (April 2021) in use with Glue.

If you're trying to create a pandas dataframe from a file stored in S3 using read_csv / read_excel, you will get this error.

You can get around it by first downloading it to a local directory and then passing that local file name to the pandas read_ function. Not pretty, I know.

Upvotes: 2

Tomasz
Tomasz

Reputation: 155

It seems that right now it is not possible to overwrite botocore and boto3 libraries versions on AWS Glue Python shell (https://github.com/boto/boto3/issues/2566).

Provided versions are:

  • botocore 1.12.232
  • boto3 1.9.203

aiobotocore is looking for some import that is not available in botocore 1.12.232.

I know that's not perfect solution, but in my case I had to remove/downgrade all dependencies that use features which are not available in those boto libraries to make Glue job work correctly.

Upvotes: 4

Related Questions