Need AWS Glue to store bad records/ records with error when reading Mongo db data to a S3 path and process the rest of the data.

Question

Hi,
Working on a solution to read Mongo db data and storing in S3. But due to conflicting schema the job does not run the second time, it always runs in the first time. I need Glue to handle the records with such conflicting schemas, tried providing **badRecordsPath** and mode = **PERMISSIVE** but it does not consider the options. Tried reading data using both -> with **catalog** and with **from_options**. Below is the snippet of the same.
1. **dynamic_frame** = glueContext.**create_dynamic_frame_from_catalog**(database = "mongo_app_config_db",table_name = "s3app_config_intf_logs",additional_options = {"database":"app_config", "collection":"intf_logs","badRecordsPath": s3_path,"mode": "PERMISSIVE" },transformation_ctx = "dynamic_frame" )
2. **dynamic_frame** = glueContext.create_dynamic_frame.**from_options**(connection_type="mongodb",connection_options=read_mongo_options, transformation_ctx = "dynamic_frame", additional_options = {"badRecordsPath": s3_path , "mode": "PERMISSIVE"})

Please help me to handle such records and continue the job run for rest of the records.

Answer

Hello,

Glue or Spark uses a connector called "**MongoDB Spark connector**" when reading from Mongodb and as per the documentation here https://www.mongodb.com/docs/spark-connector/current/configuration/read/ I do not see any options available for handling corrupted/bad records. I could also see a feature request was placed recently here https://jira.mongodb.org/browse/SPARK-327 to the community.

You can try exporting mongo collection as JSON or CSV files to an s3 location using something like https://www.mongodb.com/docs/database-tools/mongoexport/  and consume them using Spark data frames. Spark should support handling bad records for these file formats.

You can also convert between Spark data frame and Glue dynamic frame easily as shown in the below links

https://stackoverflow.com/a/59668469

https://docs.aws.amazon.com/glue/latest/dg/aws-glue-api-crawler-pyspark-extensions-dynamic-frame.html

Need AWS Glue to store bad records/ records with error when reading Mongo db data to a S3 path and process the rest of the data.

相关内容