# SerDe (Serializer/Deserializer) in Hive

Hadoop (HDFS) is designed to store raw files—it only understands bytes and does not know anything about the format, schema, or structure of the data it holds. However, Hive is a data warehouse system designed to query that data using SQL (HiveQL), meaning it expects structured rows and columns.

> Solution-

**Serialize** = convert an object/data structure into a format that can be stored or transmitted (e.g., bytes, JSON, XML).  
**Deserialize** = convert that stored/transmitted format back into a usable object/data structure.

They're inverse operations of each other.

```plaintext
import json

# Serialize: Python object -> JSON string
data = {"name": "Alice", "age": 30}
serialized = json.dumps(data)
print(serialized)        # '{"name": "Alice", "age": 30}'
print(type(serialized))  # <class 'str'>

# Deserialize: JSON string -> Python object
deserialized = json.loads(serialized)
print(deserialized)        # {'name': 'Alice', 'age': 30}
print(type(deserialized))  # <class 'dict'>
```

| Aspect | Serialize | Deserialize |
| --- | --- | --- |
| Use case | Object → String/Bytes | String/Bytes → Object |
| Common formats | JSON, XML, Protobuf, Pickle, MessagePack | Same format, reversed |

> ### UseCase-

Saving to file, sending over network, caching, reading from file, receiving from network, loading cache