Parquet Binary storing value in encoded format

Question

I'm creating a parquet file from Java using org.apache.parquet.*. No issues with other data types, but when a binary value is written and I cat the parquet file using parquet-tools, it is showing the value in encoded format. Because of that, the parquet is not processed in my system further.

Code block:

case BINARY:
recordConsumer.addBinary(stringToBinary(val));
break;
        
AND
        
private Binary stringToBinary(Object value) {
return Binary.fromString(value.toString());
}
    

Schema used is:

message m { 
required INT64 id; 
required binary username; 
required boolean active; 
}

When I cat:

parquet-tools cat <parquetFileName>

I see something like this:

id = 1
username = TmFtZTE=
active = true

id = 2
username = TmFtZTI=
active = false

I want to see the actual Username passed and not the encoded strings.

Answer 1

Try this in your schema required binary username (UTF8)