Converting Between Formats

import pandas as pd
import json

# Read Parquet file
df = pd.read_parquet('btcusdt_1d_2020-2024.parquet')

# Convert to JSON (records format - array of objects)
df.to_json('btcusdt_1d_2020-2024.json', orient='records', indent=2)

# Alternative: Split format (better for large files)
# df.to_json('btcusdt_1d_2020-2024.json', orient='split', indent=2)

print(f"Converted {len(df)} rows to JSON")
print(f"Columns: {len(df.columns)}")
print(f"File created: btcusdt_1d_2020-2024.json")

import pandas as pd

# Read JSON file
df = pd.read_json('btcusdt_1d_2020-2024.json')

# Convert to Parquet with compression
df.to_parquet(
    'btcusdt_1d_2020-2024.parquet',
    engine='pyarrow',
    compression='snappy',  # Good balance of speed and size
    index=False
)

print(f"Converted {len(df)} rows to Parquet")
print(f"Columns: {len(df.columns)}")
print(f"File created: btcusdt_1d_2020-2024.parquet")

# Check file sizes
import os
json_size = os.path.getsize('btcusdt_1d_2020-2024.json') / (1024*1024)
parquet_size = os.path.getsize('btcusdt_1d_2020-2024.parquet') / (1024*1024)
print(f"\nFile Size Comparison:")
print(f"   JSON: {json_size:.2f} MB")
print(f"   Parquet: {parquet_size:.2f} MB")
print(f"   Space saved: {((json_size - parquet_size) / json_size * 100):.1f}%")

# WINDOWS (Command Prompt or PowerShell):
python convert.py

# LINUX/MAC (Terminal):
python3 convert.py

# Or run directly (if script has shebang):
# chmod +x convert.py  # Linux/Mac only
# ./convert.py