Serialize Data Formats

Select and implement the right data serialization format for your use case, with correct encoding/decoding and performance awareness.

When to Use

• Choosing a wire format for API communication
• Persisting structured data to disk or object storage
• Exchanging data between systems written in different languages
• Optimizing data transfer size or parsing speed
• Migrating from one serialization format to another

Inputs

• Required: Data structure to serialize (schema or example)
• Required: Use case (API, storage, streaming, analytics)
• Optional: Performance requirements (size, speed, schema enforcement)
• Optional: Target language/runtime constraints
• Optional: Human readability requirements

Procedure

Step 1: Select the Right Format

Format	Human Readable	Schema	Size	Speed	Best For
JSON	Yes	Optional (JSON Schema)	Medium	Medium	REST APIs, config, broad interop
XML	Yes	XSD, DTD	Large	Slow	Enterprise/legacy, SOAP, documents
YAML	Yes	Optional	Medium	Slow	Config files, CI/CD, Kubernetes
Protocol Buffers	No	Required (.proto)	Small	Fast	gRPC, microservices, mobile
MessagePack	No	None	Small	Fast	Real-time, embedded, Redis
Arrow/Parquet	No	Built-in	Very Small	Very Fast	Analytics, columnar queries, data lakes

Decision tree:

1. Need human editing? → YAML (config) or JSON (data)
2. Need strict schema + fast RPC? → Protocol Buffers
3. Need smallest wire size? → MessagePack or Protobuf
4. Need columnar analytics? → Apache Parquet
5. Need in-memory interchange? → Apache Arrow
6. Legacy enterprise integration? → XML

Expected: Format selected with documented rationale matching use case requirements. On failure: If requirements conflict (e.g., human-readable AND fast), prioritize the primary use case and note the trade-off.

Step 2: Implement JSON Serialization

import json
from datetime import datetime, date
from dataclasses import dataclass, asdict

@dataclass
class Measurement:
    sensor_id: str
    value: float
    unit: str
    timestamp: datetime

# Custom encoder for non-standard types
class CustomEncoder(json.JSONEncoder):
    def default(self, obj):
        if isinstance(obj, datetime):
            return obj.isoformat()
        if isinstance(obj, date):
            return obj.isoformat()
        if isinstance(obj, bytes):
            import base64
            return base64.b64encode(obj).decode('ascii')
        return super().default(obj)

# Serialize
measurement = Measurement("sensor-01", 23.5, "celsius", datetime.now())
json_str = json.dumps(asdict(measurement), cls=CustomEncoder, indent=2)

# Deserialize
data = json.loads(json_str)

# R: JSON with jsonlite
library(jsonlite)

# Serialize
df <- data.frame(sensor_id = "sensor-01", value = 23.5, unit = "celsius")
json_str <- jsonlite::toJSON(df, auto_unbox = TRUE, pretty = TRUE)

# Deserialize
df_back <- jsonlite::fromJSON(json_str)

Expected: Round-trip serialization preserves all data types accurately. On failure: If a type is lost (e.g., dates become strings), add explicit type conversion in the deserialization step.

Step 3: Implement Protocol Buffers

Define the schema (.proto file):

syntax = "proto3";
package sensors;

message Measurement {
  string sensor_id = 1;
  double value = 2;
  string unit = 3;
  int64 timestamp_ms = 4;  // Unix milliseconds
}

message MeasurementBatch {
  repeated Measurement measurements = 1;
}

Generate and use:

# Generate Python code
protoc --python_out=. sensors.proto

# Generate Go code
protoc --go_out=. sensors.proto

from sensors_pb2 import Measurement, MeasurementBatch
import time

# Serialize
m = Measurement(
    sensor_id="sensor-01",
    value=23.5,
    unit="celsius",
    timestamp_ms=int(time.time() * 1000)
)
binary = m.SerializeToString()  # Compact binary

# Deserialize
m2 = Measurement()
m2.ParseFromString(binary)

Expected: Binary output 3-10x smaller than equivalent JSON. On failure: If protoc is unavailable, use a language-native protobuf library (e.g., betterproto for Python).

Step 4: Implement MessagePack

import msgpack
from datetime import datetime

# Custom packing for datetime
def encode_datetime(obj):
    if isinstance(obj, datetime):
        return {"__datetime__": True, "s": obj.isoformat()}
    return obj

def decode_datetime(obj):
    if "__datetime__" in obj:
        return datetime.fromisoformat(obj["s"])
    return obj

data = {"sensor_id": "sensor-01", "value": 23.5, "ts": datetime.now()}

# Serialize (smaller than JSON, faster than JSON)
packed = msgpack.packb(data, default=encode_datetime)

# Deserialize
unpacked = msgpack.unpackb(packed, object_hook=decode_datetime, raw=False)

Expected: MessagePack output is 15-30% smaller than JSON for typical payloads. On failure: If a language lacks MessagePack support, fall back to JSON with compression (gzip).

Step 5: Implement Apache Parquet (Columnar)

import pyarrow as pa
import pyarrow.parquet as pq
import pandas as pd

# Create data
df = pd.DataFrame({
    "sensor_id": ["s-01", "s-02", "s-01", "s-03"] * 1000,
    "value": [23.5, 18.2, 24.1, 19.8] * 1000,
    "unit": ["celsius"] * 4000,
    "timestamp": pd.date_range("2025-01-01", periods=4000, freq="min")
})

# Write Parquet (columnar, compressed)
table = pa.Table.from_pandas(df)
pq.write_table(table, "measurements.parquet", compression="snappy")

# Read Parquet (can read specific columns without loading all data)
table_back = pq.read_table("measurements.parquet", columns=["sensor_id", "value"])
df_subset = table_back.to_pandas()

# R: Parquet with arrow
library(arrow)

# Write
df <- data.frame(sensor_id = rep("s-01", 1000), value = rnorm(1000))
arrow::write_parquet(df, "measurements.parquet")

# Read (with column selection — only reads selected columns from disk)
df_back <- arrow::read_parquet("measurements.parquet", col_select = c("value"))

Expected: Parquet files 5-20x smaller than CSV for typical tabular data. On failure: If Arrow is unavailable, use fastparquet (Python) or CSV with gzip as fallback.

Step 6: Compare Performance

Run benchmarks for your specific data and use case:

import json, msgpack, time
import pyarrow as pa, pyarrow.parquet as pq

data = [{"id": i, "value": i * 0.1, "label": f"item-{i}"} for i in range(10000)]

# JSON
start = time.perf_counter()
json_bytes = json.dumps(data).encode()
json_time = time.perf_counter() - start

# MessagePack
start = time.perf_counter()
msgpack_bytes = msgpack.packb(data)
msgpack_time = time.perf_counter() - start

print(f"JSON:    {len(json_bytes):>8} bytes, {json_time*1000:.1f} ms")
print(f"MsgPack: {len(msgpack_bytes):>8} bytes, {msgpack_time*1000:.1f} ms")

Expected: Benchmark results guide format selection for production use. On failure: If performance is insufficient for any format, consider compression (zstd, snappy) as an orthogonal optimization.

Validation

• Selected format matches use case requirements (documented rationale)
• Round-trip serialization preserves all data types
• Edge cases handled: empty collections, null/None values, Unicode, large numbers
• Performance benchmarked for representative payload sizes
• Error handling for malformed input (graceful failures, not crashes)
• Schema documented (JSON Schema, .proto, or equivalent)

Common Pitfalls

• Floating-point precision: JSON represents all numbers as IEEE 754 doubles. Use string encoding for financial/decimal precision.
• Date/time handling: JSON has no native datetime type. Always document the format (ISO 8601) and timezone handling.
• Schema evolution: Adding or removing fields can break consumers. Protobuf handles this well; JSON requires careful versioning.
• Binary data in JSON: Base64 encoding inflates binary data by ~33%. Use a binary format for binary-heavy payloads.
• YAML security: YAML parsers may execute arbitrary code via !!python/object tags. Always use safe loaders.

Related Skills

• design-serialization-schema — schema design, versioning, and evolution strategies
• implement-pharma-serialisation — pharmaceutical serialisation (different domain, same naming)
• create-quarto-report — data output formatting for reports