Athena - Querying S3 file (CSV with JSON objects)

Question

Hi,
I am trying to use Athena to query the s3 file which is in csv format and contains the json objects.

**S3 CSV file:**

```
Id,  data,  last_update
001, {"key1":"value1"}, 10-01-2024
002, {"key1":"value1" , "key2":"value2"}, 10-01-2024
```

I am using comma as the delimiter and " as the quotes parameter in the classifier of glue crawler.

However, its splitting the second record's json object because of the delimiter.

**Athena query output**

```
ID       Data             LastUpdate
001   {"key1":"value1"}    10-01-2024
002   {"key1":"value1"     "key2":"value2"}
```

Can you please advise how do we handle this? Appreciate your help.

Thanks

Answer

That CSV looks broken, if you use the separator inside the field you need to escape the field.   
For instance, if you make the ' the escape character:  
002,'{"key1":"value1" , "key2":"value2"}',10-01-2024

You could work around with a custom grok classifier that knows when there are { at the beginning of a field it should consider the field only ends when it finds }

Athena - Querying S3 file (CSV with JSON objects)

Relevant content