Dear
  Thanks a lot for you create avro it's very helpful&powerful. Just as we all 
know on the one hand via hive we can use sql query data in avro file on the 
other hand avro doesn't  support the CRUD&insert operation &random access 
function. I suggest avro add the following features:
1. The CRUD&insert operation &random access function of the data included in 
the avro file.
2. Support CRUD operations &insert&random access access while supporting 
multiple processes & thread concurrency for the same avro file.
3. Performance optimization supports high-performance concurrent access to avro 
files for serialization & deserialization & CRUD&insert operations & random 
access access under multi-threaded high concurrency under NIO.
4. Support link multiplexing to perform multiple RPC interactions in the same 
TCP channel.
5. Supports the Memory Cache of the complete avro file or the data item in the 
avro file.



Avro's CRUDI&random access can of course be implemented in the form of a 
modified index (for example, changing the CRUD&insert operation to add a record 
at the end of the avro file and then updating the corresponding index). 
Additional defragmentation function to defragment when load is low (such as 
creating a new avro file, then copying the record in the original avro file to 
the new file in order and updating the index of the corresponding record, then 
rename the old file After that and change the new file's name to the old file 
and then delete the old file.


With BRs

Reply via email to